A&A 635, A152 (2020) Astronomy https://doi.org/10.1051/0004-6361/201937240 & c ESO 2020 Astrophysics

The Tucana dwarf spheroidal galaxy: not such a massive failure after all? S. Taibi1,2, G. Battaglia1,2, M. Rejkuba3, R. Leaman4, N. Kacharov4, G. Iorio5, P. Jablonka6,7, and M. Zoccali8,9

1 Instituto de Astrofisica de Canarias, C/Via Lactea s/n, 38205 La Laguna, Tenerife, Spain e-mail: [email protected], [email protected] 2 Departamento de Astrofisica, Universidad de La Laguna, 38205 La Laguna, Tenerife, Spain 3 European Southern Observatory, Karl-Schwarzschild Strasse 2, 85748 Garching, Germany 4 Max Planck Institute for Astronomy, Königstuhl 17, 69117 Heidelberg, Germany 5 Dipartimento di Fisica e Astronomia “G. Galilei”, Università di Padova, Vicolo dell’Osservatorio 3, 35122 Padova, Italy 6 Institute of Physics, Laboratory of Astrophysics, Ecole Polytechnique Federale de Lausanne (EPFL), 1290 Sauverny, Switzerland 7 GEPI, CNRS UMR 8111, Observatoire de Paris, PSL Research University, 92125 Meudon Cedex, France 8 Instituto de Astrofisica, Pontificia Universidad Catolica de Chile, Av. Vicuña Mackenna 4860, 782-0436 Macul, Santiago, Chile 9 Millennium Institute of Astrophysics, Av. Vicuña Mackenna 4860, 782-0436 Macul, Santiago, Chile Received 3 December 2019 / Accepted 28 January 2020

ABSTRACT

Context. Isolated local group (LG) dwarf galaxies have evolved most or all of their life unaffected by interactions with the large LG spirals and therefore offer the opportunity to learn about the intrinsic characteristics of this class of objects. Aims. Our aim is to explore the internal kinematic and metallicity properties of one of the three isolated LG early-type dwarf galaxies, the Tucana dwarf spheroidal. This is an intriguing system, as it has been found in the literature to have an internal rotation of up to 16 km s−1, a much higher velocity dispersion than dwarf spheroidals of similar luminosity, and a possible exception to the too-big- too-fail problem. Methods. We present the results of a new spectroscopic dataset that we procured from the Very Large Telescope (VLT) taken with the FORS2 instrument in the region of the Ca II triplet for 50 candidate red giant branch stars in the direction of the Tucana dwarf spheroidal. These yielded line-of-sight (l.o.s.) velocity and metallicity ([Fe/H]) measurements of 39 effective members that double the number of Tucana’s stars with such measurements. In addition, we re-reduce and include in our analysis the other two spectroscopic datasets presented in the literature, the VLT/FORS2 sample by Fraternali et al. (2009, A&A, 499, 121), and the VLT/FLAMES one from Gregory et al. (2019, MNRAS, 485, 2010). Results. Across the various datasets analyzed, we consistently measure a l.o.s. systemic velocity of 180 ± 1.3 km s−1 and find that a dispersion-only model is moderately favored over models that also account for internal rotation. Our best estimate of the internal l.o.s. +1.6 −1 velocity dispersion is 6.2−1.3 km s , much smaller than the values reported in the literature and in line with similarly luminous dwarf spheroidals; this is consistent with NFW halos of circular velocities <30 km s−1. Therefore, Tucana does not appear to be an exception to the too-big-to-fail problem, nor does it appear to reside in a dark matter halo much more massive than those of its siblings. As for the metallicity properties, we do not find anything unusual; there are hints of the presence of a metallicity gradient, but more data are needed to pinpoint its presence. Key words. galaxies: dwarf – Local Group – galaxies: stellar content – galaxies: kinematics and dynamics – galaxies: abundances – techniques: spectroscopic

1. Introduction the MW, obtained by integrating, back in time, their present- day systemic motion (see e.g., Gaia Collaboration 2018; Fritz Dwarf galaxies are the least massive, yet the most dark- et al. 2018; Simon 2018, for Gaia-DR2 based determinations), matter-dominated galactic systems observed (e.g., Mateo 1998; are consistent with repeated pericentric passages for several of Battaglia et al. 2013; Walker 2013). In the Local Group (LG), these objects. In practice, (unknown) factors such as the triax- the nearest ones to the largest spirals, meaning the Milky Way iality of the MW’s potential, or interactions between satellites, (MW) and M 31, are gas-poor dwarf spheroidal systems (dSphs) among others, introduce significant uncertainties in the recon- with no on-going star formation (Tolstoy et al. 2009). Although struction of their full orbital history (e.g., Lux et al. 2010), but a they share the same morphology, their full star formation histo- comparison with the properties of dark matter sub-halos around ries show complex evolutionary pathways (Gallart et al. 2015). MW-sized hosts does suggest that most MW satellites fell in Due to their small masses, the formation and evolution of the MW halo at intermediate to early times (see Rocha et al. these galaxies could be strongly influenced by environmental 2012; Wetzel et al. 2015). A large body of work has therefore effects. The orbital properties of the dwarf galaxy satellites of focused on exploring the impact of tidal and/or ram-pressure ? Based on observations made with ESO telescopes at the La Silla stripping caused by a MW-sized host onto various properties Paranal Observatory as part of the programs 091.B-0251, 69.B-0305(B) of the dSph satellites of the MW (see e.g., Piatek & Pryor 1995; and 095.B-0133(A). Mayer et al. 2001a,b, 2006; Read et al. 2006; Muñoz et al. 2008;

Article published by EDP Sciences A152, page 1 of 24 A&A 635, A152 (2020)

Klimentowski et al. 2009; Kazantzidis et al. 2011; Pasetto et al. well-described by an exponential fit. The recovery of the full 2011; Battaglia et al. 2015; Iorio et al. 2019). However, as was star formation history (SFH) from deep HST/ACS observations recently shown by Hausammann et al.(2019), the ram-pressure reaching the oldest main sequence turn-off by Monelli et al. stripping induced by a host halo has its limits in the actual (2010) showed that Tucana formed the majority of its stars more quench and gas depletion of a dSph. Internal effects, like stellar than 9 Gyr ago. It experienced a strong initial period of star feedback resulting from episodic star formation and supernova- formation (SF) starting very early on (∼13 Gyr ago). Tucana driven winds, play an important role too (see e.g., Sawala et al. harbors at least two stellar sub-populations based on observed 2010; Bermejo-Climent et al. 2018; Revaz & Jablonka 2018). splitting of the HB, double red giant branch (RGB) bump, and Recent hydro-dynamic cosmological simulations have indeed the luminosity-period properties of the RR-Lyrae, which imply shown that both internal and environmental mechanisms are nec- that this system experienced at least two early phases of SF in a essary to the reproduction of the observed properties of the LG short period of time. Using the same HST/ACS dataset, Savino dwarf galaxies (see e.g., Brooks & Zolotov 2014; Garrison- et al.(2019) refined the HB analysis, showing that Tucana expe- Kimmel et al. 2014; Wetzel et al. 2016; Sawala et al. 2016). rienced two initial episodes of sustained SF followed by a third Stellar feedback, for example, was also found to be an important less intense, but more prolonged one, ending between 6 and ingredient in enhancing tidal-stirring effects (Kazantzidis et al. 8 Gyr ago. The spatial analysis of the same dataset indicates the 2017). presence of a population age gradient inside ∼4 Re (Monelli et al. Given the multitude of physical mechanisms affecting low- 2010; Hidalgo et al. 2013; Savino et al. 2019). mass galaxy evolution, data on the ages, chemical abundances, The first spectroscopic study of individual stars in the spatial distribution, and kinematics of the stellar component of Tucana dSph was conducted by Fraternali et al.(2009), with the LG dwarf galaxies are needed to understand the observed diver- VLT/FORS2 instrument obtaining a relatively small sample of sity of these systems. Detailed observations of MW satellites ∼20 RGB probable member stars. They reported the systemic −1 have been accumulated over the past years including large spec- velocity and velocity dispersion values (¯vsys = 194.0±4.3 km s troscopic datasets (e.g., Tolstoy et al. 2004; Battaglia et al. 2006, +4.1 −1 and σv = 15.8−3.1 km s ) for the galaxy, together with the pres- 2008, 2011; Walker et al. 2009a; Kirby et al. 2011; Lemasle ence of a maximum rotation signal of ∼16 km s−1. They also et al. 2012, 2014; Hendricks et al. 2014; Spencer et al. 2017). determined a mean metallicity value of [Fe/H] = −1.95 ± The observations showed little to no sign of internal rotation 0.15 dex with a dispersion of 0.32 ± 0.06 dex. The determination (e.g., Battaglia et al. 2008; Walker et al. 2008; Leaman et al. of the systemic velocity ruled out an association with a nearby 2013; Wheeler et al. 2017), pronounced radial metallicity gradi- HI cloud, confirming that Tucana is devoid of neutral gas (down 4 ents (e.g., Battaglia et al. 2006, 2011; Walker et al. 2009a; Kirby to a HI mass of 1.5 × 10 M ), and in addition it is moving away et al. 2011; Leaman et al. 2013) and multiple populations with from the LG barycenter (Fraternali et al. 2009). If bound, Tucana different chemo-dynamical properties (e.g., Tolstoy et al. 2004; has not reached its apocenter yet. It is indeed possible that a Battaglia et al. 2006, 2008). Nonetheless, very few simulations past interaction between Tucana and the MW happened around have explored the link between the formation scenarios and the 10 Gyr ago (roughly coinciding with the major drop in its SF; internal kinematic and chemical status of dwarf galaxies (Revaz Sales et al. 2007; Fraternali et al. 2009; Teyssier et al. 2012). et al. 2009; Schroyen et al. 2013; Revaz & Jablonka 2018). A new spectroscopic study of Tucana’s RGB stars has recently It is likely that the MW influenced at least in part the evo- been conducted by Gregory et al.(2019) using VLT /FLAMES lution of its satellites. Therefore, isolated dSphs represent a tool data. Their sample of probable members is slightly larger than that is valuable in the understanding of the intrinsic properties that of Fraternali et al.(2009), but covers a much more extended of the systems that have spent all or most of their life in a more spatial area (up to ∼10 R ∼ 2 R ). Their velocity dispersion benign environment. They are crucial to our understanding of e tidal value (σ = 14.4+2.8 km s−1) obtained for 36 probable members the possible formation scenarios. In the LG, just a handful of v −2.3 is similar to that of Fraternali et al.(2009), although their sys- dSphs are found in isolation: namely And XVIII, Cetus, and temic velocity shows a significant offset (¯v = 216.7+2.9 km s−1); Tucana. Although it is probable that in the past they may have sys −2.8 +4.2 −1 −1 interacted just once with the MW or M 31 based on their line-of- they also detect a velocity gradient of k = 7.6−4.3 km s arcmin sight (l.o.s.) systemic velocity (Lewis et al. 2007; Fraternali et al. along the optical major axis. Performing a dynamical model- 2009; but see also Sales et al. 2010; Teyssier et al. 2012), the fact ing of Tucana’s kinematic properties based on the FLAMES/ that they spent most of their lives in isolation makes them ideal GIRAFFE l.o.s. velocities of the probable members, the authors targets to contrast with the satellite dwarfs. found a massive dark matter halo with a high central density. This work is part of a larger body of studies aiming to The implied dark matter halo mass profile is much denser than improve our knowledge of the observed properties of a selected the other dwarfs of the LG, making Tucana the first exception sample of isolated LG dwarf galaxies – Phoenix, (Kacharov et al. of the too-big-to-fail problem (see e.g., Boylan-Kolchin et al. 2017); Cetus, (Taibi et al. 2018); Aquarius, (Hermosa Muñoz 2012). In fact, the pure N-body simulations in the framework et al. 2020) – mainly exploiting VLT/FORS2 multi-object spec- of Λ cold dark matter (Λ-CDM) predict that the dwarf galax- troscopic data. Here we focus on the Tucana dSph. ies of the LG should live in denser halos than those inferred Tucana is an early-type dwarf galaxy found in extreme isola- from observations. The fact that Tucana is found to reside in tion with a heliocentric distance of D = 887 ± 49 kpc (Bernard such a massive halo in agreement with Λ-CDM predictions et al. 2009), which places it at more than 1 Mpc away from seems to indicate that, during its evolution, it has been able M 31 and the LG center, with only the Phoenix dwarf found to maintain its initial content of dark matter, independently of within ∼500 kpc of it (McConnachie 2012). Photometric obser- the internal and environmental mechanisms that have driven its vations have shown that the galaxy is mainly old and metal-poor, evolution. with an extended horizontal branch (HB) and a population of Motivated by the unique internal kinematics of Tucana and variable stars (see e.g., Saviane et al. 1996; Bernard et al. the importance of isolated dwarfs in disentangling evolutionary 2009). The structural analysis by Saviane et al.(1996) showed processes, in this study we present results from a new investi- a highly flattened system (e ∼ 0.5) with a surface density profile gation of the kinematic and chemical properties of the stellar

A152, page 2 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph

Table 1. Parameters adopted for the Tucana dSph. 2.1. The P91 FORS2 dataset The primary dataset analyzed in this work was obtained with Parameter Units Value Ref. the FORS2 instrument mounted at the UT1 (Antu) of the h m s Very Large Telescope (VLT) at the ESO Paranal observatory. αJ2000 22 41 49.6 (1) ◦ 0 00 Observations were taken in service mode over several nights δJ2000 −64 25 10 (1)  (a) 0.48 ± 0.03 (2) between July 2013 and July 2014 as part of the ESO program PA deg 97 ± 2 (2) 091.B-0251, PI: M. Zoccali (see Table2). The instrument was R arcmin (pc) 0.7 ± 0.1 (955 ± 139) (2) used in multi-object spectroscopic mode (MXU), which allows core the observer to employ exchangeable masks with custom-cut Rtidal arcmin (pc) 3.7 ± 0.5 (181 ± 28) (2) R arcmin (pc) 0.8 ± 0.1 (206 ± 28) (2) slits. We used pre-imaging FORS2 photometry taken in John- e V I L 105 M 5.5 ± 1.5 (2) son - and -band, to allocate slits to stars with colors and V magnitudes compatible with Tucana’s RGB. Slits that would oth- ITRGB 20.7 ± 0.15 (2) E(B − V) 0.031 (3) erwise have remained empty (five of them) were allocated to ran- D kpc 887 ± 49 (3) dom targets in the same magnitude range. We selected 50 objects distributed over two overlapping masks of 27 slits each; the v¯ km s−1 180 ± 1.3 (4) sys observation of four objects was repeated on purpose for internal σ km s−1 6.2+1.6 (4) v −1.3 accuracy measurements. Therefore, the selected targets covered [Fe/H] dex −1.58 (4) an area up to the nominal King tidal radius of Tucana (Rtidal = σ[Fe/H] dex 0.39 (4) 3.70), as can be seen in Fig.1. Notes. The table lists: the coordinates of the galaxy’s optical center; the The adopted instrumental set-up and observing strategy were ellipticity; the position angle; the core, tidal and half-light geometric the same as in our previous studies (Kacharov et al. 2017; Taibi radii; the stellar luminosity in V-band; the tip of the red giant branch et al. 2018; Hermosa Muñoz et al. 2020), so we only report magnitude in I-band; the average reddening; the heliocentric distance; the essentials here. We used the 1028z+29 holographic grism the chemo-kinematic parameters obtained in this work, i.e., the systemic together with the OG590+32 order separation filter in order to velocity, the velocity dispersion, the median metallicity, and the intrin- (a) cover the wavelength range between 7700 and 9500 Å. Slits had sic metallicity scatter.  = 1 − b/a. spatial sizes of 100 × 1000 (800 in some cases to avoid overlaps) References. (1) Lavery & Mighell(1992); (2) Saviane et al.(1996); (3) in the first mask (Tuc0) and of 100 × 800 (700 for overlaps) in Bernard et al.(2009); (4) this work. the second one (Tuc1). This led to a binned spectral dispersion −1 of 0.84 Å pxl and a resolving power of R = λcen/∆λ ∼ component of the Tucana dSph. We analyzed a new dataset of 2600 at λcen = 8600 Å (equivalent to a velocity resolution of multi-object spectroscopic observations of 50 individual RGB 28 km s−1 pxl−1). Ten identical observing blocks (OBs1) were stars taken with the VLT/FORS2 instrument targeting the near- taken for each pointing in order to reach the necessary signal- IR wavelength region of the Ca II triplet (CaT) lines. To under- to-noise ratio (S/N) for velocity and metallicity measurements. stand how systematics may influence the key results of our In Table2, we report a complete observing log. study, we further re-reduced the original datasets presented in The data were provided by ESO as individual OB-datasets Fraternali et al.(2009) and Gregory et al.(2019) and performed within the FORS2 standard delivery plan. We adopted the a combined analysis together with our own data. In this work, we same data-reduction process as in Taibi et al.(2018, hereafter, present an in-depth and homogeneous analysis of all currently T18), based on IRAF2 routines and custom-made python scripts. available spectroscopic data for the Tucana dSph. Briefly, our pipeline was developed to organize and reduce each The article is structured as follows: in Sect.2, we present the OB-dataset independently. After making bias and flat-field cor- data acquisition and reduction processes for all the datasets ana- rections on the two-dimensional (2D) multi-object scientific and lyzed in this work. Section3 is dedicated to the determination lamp-calibration frames, these were also cleaned from cosmic of the l.o.s. velocity and metallicity measurements. In Sect.4, rays and bad rows. The 2D images are corrected for distortions in we describe the criteria applied to select likely member stars the spatial direction by rectifying their slit traces in order to cut in the different datasets we present here. Section5 shows the them into individual 2D spectra. The arc-lamp spectra are then results from the kinematic analysis, presenting the determination used to find the wavelength solution to calibrate the scientific of the galaxy systemic velocity and velocity dispersion, along exposures, with a typical RMS accuracy of 0.05 Å. The wave- with the search for a possible rotation signal and the implica- length calibration also has the important effect of rectifying the tion for the dark matter halo properties of Tucana. In Sect.6, sky lines, which had initially been curved by the instrument dis- we describe the determination of metallicities ([Fe/H]) and the perser, helping to reduce the residuals during the sky-subtraction subsequent chemical analysis. Finally, Sect.7 is dedicated to our part. The rectified wavelength-calibrated 2D individual scien- summary and conclusions, while in the appendices we report the tific exposures were finally background subtracted, optimally detailed comparison between the measurements obtained for our extracted into 1D spectra and normalized by fitting the stellar dataset with those reported in the literature, along with supple- mentary material from the kinematic analysis. The parameters adopted for the Tucana dSph are summarized in Table1. 1 The observations are organized by ESO in blocks taking into account the time to actually spend on the object, including foreseen overheads. Data are later delivered as OB-datasets which include the scientific 2. Data acquisition and reduction processes exposures related to the individual OBs together with the associated acquisition and standard calibration frames (biases, arc lamp, dome flat The analysis we present in this work required combining data fields). obtained with different instruments and observational set-ups. 2 IRAF is the Image Reduction and Analysis Facility distributed by the Here, we provide details about the data acquisition and reduc- National Optical Astronomy Observatories (NOAO) for the reduction tion processes we followed. and analysis of astronomical data: http://iraf.noao.edu/

A152, page 3 of 24 A&A 635, A152 (2020)

Table 2. Observing log of the P91 VLT/FORS2 MXU observations of RGB targets along the line of sight to the Tucana dSph.

Field Position (RA, Dec) Date/time Exp. Airmass DIMM seeing Grade (?) Slits (J2000) (UT) (s) (00) Tuc0 22:41:56, −64:24:43 2013-07-30/04:26 3400 1.44 0.78 B 27 (16+11) 2013-08-01/04:46 3400 1.39 0.80 A 2013-08-01/05:44 3400 1.32 0.92 A 2013-08-01/06:42 3400 1.30 0.90 A 2013-08-05/06:04 3400 1.30 0.98 A 2013-08-05/07:02 3400 1.31 1.13 A 2013-08-13/03:49 3400 1.41 0.69 A 2013-08-13/04:47 3400 1.33 0.57 A 2013-08-28/03:18 3400 1.36 0.65 A 2013-08-30/03:43 3400 1.33 0.89 A Tuc1 22:42:04, −64:24:47 2013-09-05/02:07 3500 1.43 1.62 C (a) 27 (20+7) 2013-09-06/03:34 3500 1.31 0.77 A 2013-09-06/04:37 3500 1.30 0.89 A 2013-09-13/01:53 3500 1.40 1.49 C (b) 2013-09-13/02:54 3500 1.32 1.45 B 2013-10-01/01:35 3500 1.33 0.68 A 2013-10-25/02:17 3500 1.33 1.04 B 2014-07-21/07:28 3500 1.30 0.92 A 2014-07-21/08:40 3500 1.33 1.05 A 2014-07-27/06:07 3500 1.32 0.70 A 2014-07-27/07:10 3500 1.30 0.60 A 2014-07-27/08:18 200 1.33 0.65 C (c) 2014-07-29/05:14 3500 1.37 0.62 A Total 54

Notes. From left to right, column names indicate: the pointing field name; the field center coordinates; observing date and starting time of the scientific exposure; the exposure time in seconds; the starting airmass; the average DIMM seeing during the exposure in arcsec; the ESO OB fulfillment grades (a full description is reported in the notes below); the number of slits/observed objects per mask. For each field, the mask design remained identical in each OB. The total number of slits (54) is reported in the last row of the table. (?)ESO OB fulfillment Grades: (A) fully within constraints – OB completed; (B) mostly within constraints, some constraint is 10% violated – OB completed; (C) out of constraints – OB must be repeated: (a)seeing increased up to ∼1.200 at OB’s end – FWHM of spectra ∼1.000. (b)Seeing, FLI, and Moon distance out of constraints. (c)Aborted after 200 s because sky conditions changed to thin clouds.

Fig. 1. Spatial distribution (left) and color-magnitude diagram (right) of stars along the line of sight to the Tucana dSph. Black points represent the objects classified as stars in the FORS2 photometric data (see main text); red and blue dots indicate the P91-FORS2 MXU targets classified as probable members (i.e., with P > 0.05) and non-members, respectively. Yellow triangles and green squares represent the probable member stars from the P69-FORS2 and FLAMES datasets, respectively, which were added to the P91-FORS2 and analyzed through the text. The two observed FORS2 pointings are represented as large squares, while the ellipses denote the galaxy half-light radius and the spatial extension of the dataset (i.e., up to 5 × Re ∼ Rtidal). We note that the photometric data are not corrected for reddening.

A152, page 4 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph continuum. The median S/N around the CaT for the individual The final P69 sample was reduced to 28 objects, with exposures was ∼8 Å−1. five more than the initially published F09 catalog. Once We then followed the approach already presented in T18 stacked together, reliable spectra had an S/N ∼ 22 Å−1 (see also and Hermosa Muñoz et al.(2020) to stack the repeated indi- Table C.2 for the properties of each individual observed star). vidual exposures together for each target. To do so, we needed Another issue was the lack of through-slit images in the pro- to account for possible zero-point displacements in the wave- vided data, although in F09 it is reported that the objects were length calibration, small slit-centering shifts, and the different well-centered in the spectral direction after visual inspection dur- dates of observation. We used the numerous OH emission lines ing the spectroscopic run. However, we took into account the in the extracted sky background to refine the wavelength cali- error related to the slit centering by assuming that it is a tenth of bration of the individual spectra, using the IRAF fxcor task to a pixel (∼3 km s−1, which is the typical error found checking for cross-correlate with a reference sky spectrum over the region the slit centering) and adding it in quadrature during the velocity 8250−9000 Å. The calculated off-sets roughly varied between estimation step. 5 and 25 km s−1 with an average error of 2 km s−1. The cor- As can be seen in Appendix A.2, there is good agreement for rection for slit-centering shifts was done using the through-slit the measurements of the stars in common between this dataset frames typically taken before each scientific exposure. The off- (the P69 FORS2) and the P91 FORS2 (and for those cases where sets were calculated as the difference in pixels between the slit there is disagreement, the source of it can be traced back). center and the star centroid for every target per mask. The cor- rection for each target was taken as the median value of all its slit shifts; the associated error was the scaled median absolute 2.3. The FLAMES dataset deviation (MAD) of those values. We found slit shifts in the Gregory et al.(2019, hereinafter, G19) recently presented an range ±0.1−9.5 km s−1 with errors of ±1−5 km s−1. Finally, we additional sample of l.o.s. velocities (not metallicities) for obtained the heliocentric correction for all the individual expo- individual stars in the direction of Tucana, taken with the sures using the IRAF rvcorrect task. All the above shifts were FLAMES/GIRAFFE instrument at the VLT, as part of the ESO applied to the individual spectra using the IRAF dopcor task. program 095.B-0133(A), PI: M. Collins. The spectrograph was We then regrouped and averaged the repeated exposures of each used in MEDUSA mode, meaning in multi-fiber configuration, target, weighting them with their associated σ-spectra, given which allows for the simultaneous observation of up to 132 sep- by the extraction procedure. Final error spectra were obtained arate targets (sky fibers included). The instrument field of view accordingly. The median S/N around the CaT for the stacked (FoV) has a 25 arcmin diameter, and each fiber has an aperture spectra is ∼27 Å−1 (see also Table C.1 for the properties of each on the sky of 1.2 arcsec. The grating used was the LR8, centered individual observed star). on 8817 Å and covering the CaT wavelength region, yielding a spectral resolution of R ∼ 6500. 2.2. A new reduction for the Fraternali et al.(2009) FORS2 The authors reported the detection of 36 probable member dataset stars, out to very large distances from Tucana’s center, approx- imately up to 10 half-light radii. Given the larger spatial region Fraternali et al.(2009, hereinafter, F09) presented l.o.s. veloc- probed by these data with respect to the FORS2 P91 and P69 ities and metallicities for 23 individual stars with magnitudes datasets (compare Fig.1 of this work to Fig. 1 in G19), it was and colors compatible with Tucana’s RGB (of which 17 were interesting to explore whether the G19 catalog of l.o.s. veloc- classified as members) from an earlier FORS2 MXU dataset ities could be used together with our determinations from the (ESO program 69.B-0305(B), PI: E. Tolstoy). We wished to use P91 and P69 FORS2 data. The comparison of the six stars in this catalog to increase our sample size; however, a comparison common between the P91 and G19 samples yields a discrepancy with the l.o.s. velocities and metallicities between the targets in of ∼30 km s−1 for five of the six stars and of about −150 km s−1 common (3) with the F09 catalog showed significant system- for the other object, which does not allow the direct combina- atic shifts in both quantities (see Appendix A.3 for details). For tion of the measurements. The G19 paper also reports an offset the sake of homogeneity, we therefore reduced the F09 dataset of ∼23 km s−1 between their velocities and those in the F09 cat- again following the same procedure as adopted for the P91 sam- alog. We note that the offset of ∼30 km s−1 is compatible with ple. Hereinafter, we refer to this additional sample as the P69 the 23 km s−1 offset between G19 and F09 and the ∼7 km s−1 FORS2 dataset. offset we found when comparing our P91 velocities to the F09 The instrumental set-up of the P69 program was similar to catalog (see Appendix A.3 for further details). Given the above the one we adopted for our observations, with the difference information, we proceeded to perform our own reduction of being that the size of each slit was of 1.200 × 800, which translates the FLAMES/GIRAFFE data, the characteristics of which we into a slightly lower spectral resolution (≈30 km s−1 pxl−1). A briefly describe below. total of 45 initial targets were assigned to an equivalent number The observations were taken on six nights spread between of slits packed into a single pointing centered on Tucana. The June and September 2015, using two different fiber setups cov- total exposure time of the observations was 5.2 h. ering the same area: the first one (Tuc-1) had a total of 7 h The reduction of some slits, in particular the background of exposure time accumulated over seven OBs, and the sec- subtraction step, turned out to be particularly problematic, since ond setup (Tuc-2) got 6 h of exposure time taken within six these targets were not well-centered along the slit spatial direc- OBs, of which one was repeated twice. Each OB consisted of tion, but placed at their edges. This led to 15 extracted spec- 3 × 1200 s exposures. The total number of individual targets was tra with high sky residuals, which made them unreliable for 164. The first setup had 14 fibers, and the second had 15 fibers l.o.s. velocity and metallicity estimates. In addition, these targets allocated to empty sky regions distributed over the entire FoV. had magnitudes and colors outside the RGB of Tucana. There- Only very few targets are spatially found inside the tidal radius fore, we excluded them from the sample, along with two further of Tucana (∼30 targets), with the others scattered over an area objects whose extracted spectra did not show any CaT lines. much larger than the nominal extension of this galaxy. In fact,

A152, page 5 of 24 A&A 635, A152 (2020) the two pointings are off-centered by ∼90 from the optical center 2.4. Photometric data of Tucana. Photometric data were used for all of the three datasets (FORS2 The FLAMES data were downloaded from the science por- P91 and P69, and FLAMES/GIRAFFE) to exclude those objects tal of the ESO archive as already processed spectra, meaning whose magnitude and color were not compatible with being stars pre-reduced, wavelength-calibrated, extracted, and corrected to on Tucana’s RGB (see Sect.4), and for the FORS2 datasets to the barycentric velocity, but with no sky subtraction applied. For determine metallicities from the equivalent width of the CaT each OB, ESO delivers spectra stacked at the OB-level for the lines using calibrations from the literature (see Sect.3). science targets and several auxiliary data, including the individ- We used the pre-imaging FORS2 photometric catalog intro- ual scientific exposures within an OB for both scientific targets duced in Sect. 2.1 to associate V- and I-band magnitudes to the and sky fibers3. The only remaining steps were to perform the target stars in the P91 and P69 spectroscopic datasets. The pho- sky subtraction, combine the spectra with repeated exposures, tometric catalog was astrometrized and the instrumental magni- and, finally, calculate the radial velocities. tudes were calibrated with the publicly available catalog from We first verified the quality of the wavelength calibration by Holtzman et al.(2006) obtained with the Wide Field and Plane- cross-correlating the sky lines of the scientific exposures with tary Camera 2 of the Hubble Space Telescope (HST/WFPC2) a template sky spectrum, which was placed at the rest frame being used as a reference. We used the aperture-photometry and obtained with a similar observational set-up (Battaglia et al. catalog provided in Johnson’s UBVRI-system to actually find 2011)4. However, the delivered spectra are shifted to the helio- the astrometric solution of the FORS2 catalog and to perform centric frame, including the sky lines, thus we had to remove this the photometric calibration using the suite of codes CataXcorr correction first. We then performed the cross-correlation for each and CataComb, kindly provided to us by P. Montegriffo and spectrum of each OB dataset using the IRAF fxcor task, obtain- M. Bellazzini (INAF-OAS). ing median offsets around 0.6 km s−1 with a global scatter of The case of the FLAMES dataset, on the other hand, was dif- 0.7 km s−1. Therefore, the uncertainty related to the wavelength ferent. Since we did not have a photometric catalog covering an calibration resulted well below those from the velocity measure- area as wide as that of the spectroscopic targets, we instead used ment (as shown later). the publicly available photometry of individual point sources For the sky subtraction, we used the ESO skycorr tool (Noll from the first data release of the Dark Energy Survey (DES- et al. 2014). The idea behind this code is to adopt a physically DR1, Abbott et al. 2018). We found a match for 154 out of 164 motivated group scaling of the sky emission lines with respect to spectroscopic targets, considering a tolerance radius of 1 arcsec. a reference sky spectrum according to their expected variability To be conservative, we did not exclude the targets that did not and given the date of the observations. We created the reference have a match in the DES-DR1 photometry from the photomet- sky spectrum by first median combining the spectra of the fibers ric selection. The DES-DR1 griz-photometry needed then to be allocated to sky within the individual sub-exposures of each OB, converted first to the SDSS griz-system5 and finally to John- and then by median combining the results for the individual sub- son’s system (see Jordi et al. 2005). The FLAMES data showed exposures. The optimized line groups in the reference sky spec- magnitudes as high as V ∼ 23, which was also the limit of trum are scaled to fit the emission lines in the science spectra the DES-DR1 catalog, and, consequently, some targets had rela- and finally subtracted together with the sky continuum. Error tively large associated magnitude errors (δ ∼ 0.2). spectra are also an output of the code. We used default input mag parameter while running skycorr. This method yielded satisfac- tory results for the majority of stars, particularly for spectra with 3. Line-of-sight velocity and metallicity low S/N. measurements The sky-subtracted science spectra were then normalized using a Chebyshev polynomial of order 3, together with their The determination of l.o.s. velocities and metallicities ([Fe/H]) error spectra. Finally, repeated exposure of individual targets from the stacked spectra was done in the same way as in T18. (including those the Tuc-1 and Tuc-2 set-ups have in common) While we were able to measure both l.o.s. velocities and metal- were stacked together using a weighted average, with their error licity for the FORS2 P91 and P69 data, only l.o.s. velocities spectra combined accordingly. The typical S/N was ∼11 Å−1, could be determined for FLAMES/GIRAFFE data, due to the although in some cases it was as low as ∼2 Å−1 (see also Table C.3 low S/N of those spectra. for the properties of each individual observed star). Line-of-sight velocities were obtained using the fxcor task The average S/N measured in our reduction of the by cross-correlating with a synthetic spectrum resembling a low- FLAMES/GIRAFFE spectra is in good agreement with the S/N metallicity RGB star convolved at the same spectral resolution obtained from the GIRAFFE Exposure Time Calculator (ETC) of the dataset under consideration. For FORS2 we used the same considering typical values from the observed dataset: we used template as in T18, and cross-correlated in the wavelength range a black body template of Teff = 4500 K, an I-band magnitude 8400−8700 Å. For the FLAMES dataset we used a template of 21, an airmass of 1.2, a moon illumination fraction of 0.2, a from Zoccali et al.(2014), that is a synthetic spectrum of a star 00 00 seeing of 1.0 , an object-fiber displacement of 0.3 and a total with Teff = 4750 K, log(g) = 2.5 and [Fe/H] = −1.3 dex. In this exposure time of 10 h (36 000 s). With this setting we obtained a case, we used the region around the two reddest lines of the CaT calculated S/N of 15 Å−1, close to our typical S/N value. for the cross-correlation, since the first line often suffered from high residuals left by the subtraction of a sky emission line. This was not necessary for the FORS2 spectra, which suffer less from 3 See http://www.eso.org/observing/dfo/quality/PHOENIX/ GIRAFFE/processing.html for details. this problem due to their higher S/N. 4 We further checked the wavelength calibration of the template spec- In the following, we only keep objects whose stacked FORS2 −1 trum itself using a high-resolution atlas of sky emission lines taken with and FLAMES spectra have a S/N & 10 Å , since this is the the VLT/UVES instrument (Hanuschik 2003), degraded to the spec- limit where the velocity errors provided by fxcor task appear tral resolution of the template spectrum. The cross-correlation between these two spectra showed no significant wavelength shift. 5 https://des.ncsa.illinois.edu/releases/dr1/dr1-faq

A152, page 6 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph

G19 FLAMES 20 F09 20 P69-FORS2 P91-FORS2 P91-FORS2

15 15 N N 10 10

5 5

0 0 100 50 0 50 100 150 200 250 300 100 50 0 50 100 150 200 250 300 Vhel [km/s] Vhel [km/s] Fig. 2. Histogram of l.o.s. velocity measurements from the analyzed datasets. Left: comparing the velocities from the P91-FORS2 dataset with those from literature, i.e. Fraternali et al.(2009) and Gregory et al.(2019). Right: same comparison using instead the velocity measurements from our reduction of the P69-FORS2 and FLAMES datasets. Note that, in the left panel, the peaks of the histograms fall at different velocities, while in the right panel, where the datasets shown were analyzed homogeneously, these differences are absent. reliable (see tests in Appendix A.1). Average velocity errors tage = 8 Gyr and [Fe/H] ∼ −0.4 dex to fix the blue and red color for the FORS2 (FLAMES) spectra were found to be ∼7 km s−1 limits on the color-magnitude diagram (CMD). This color range (∼9 km s−1). was chosen to broadly cover the expected range of metallici- As can be seen from the histogram in the right panel Fig.2, ties and stellar ages obtained from the SFH analysis of Tucana the three datasets show a clear peak in the l.o.s. velocities around (Monelli et al. 2010; Savino et al. 2019). For the FLAMES 180 km s−1, where we roughly expect to find the members of dataset, due to the larger errors in the associated photometric Tucana. The FLAMES dataset presents a higher fraction of con- dataset, we broadened the blue and red color limits applied to taminants due to the large area covered by the observations. With the FORS2 case by 0.2 mag. We further excluded the targets our homogeneous analysis, we show a clear detection of the stars with spectra that showed high sky-residuals (and with S/N val- with velocities compatible with Tucana in all three datasets and ues <10 Å−1, as previously discussed) from all catalogs. Our P91 no significant offsets, unlike what we found in direct comparison sample targets reduced from 50 to 43, the P69 objects reduced with our velocity measurements and those published in F09 and from 28 to 23, and the FLAMES one went from 164 to 58. G19, as shown in the left panel of Fig.2. We refer the reader to Given that the target selection of the three datasets was car- Appendix A.3 and Sect. 5.2 for a detailed comparison with the ried out completely independently, and might therefore include studies taken from literature. different biases, we proceeded by determining memberships To estimate the [Fe/H] values, we adopted the Starkenburg and kinematic parameters by first considering our P91 FORS2 et al.(2010) relation, which is a function of the equivalent widths dataset on its own, a second time by combining it with the P69 (EWs) of the two reddest CaT lines, and of the (V − VHB) sample, and, finally, while considering all the three sets. When term, where VHB is the mean magnitude of the galaxy’s hori- we found common targets among the combined catalogs, we first zontal branch (HB). We obtained the EWs from the continuum- kept those from the FORS2 datasets, with a higher priority for normalized stacked spectra by fitting a Voigt profile over a win- those of P91. Therefore, the combined FORS2 dataset has 63 dow of 15 Å around the CaT lines of interest, and using the corre- targets, while including the FLAMES data gave a total of 109. sponding error-spectra as the flux uncertainty at each pixel in the We then continued by assigning a membership probability fitting process. The errors on the EWs were then calculated from to the individual targets in each considered dataset, applying a the covariance matrix of the fitting parameters. For the VHB, we method based on the expectation maximization technique out- adopted the value of 25.32 from Bernard et al.(2009). Uncertain- lined in Walker et al.(2009b), but with the few modifications ties on the [Fe/H] values were obtained by propagating the errors introduced by Cicuéndez et al.(2018). Briefly speaking, this on the EWs, accordingly. Typical [Fe/H] errors were found to be approach makes it possible to carry out a Bayesian analysis to around 0.15−0.25 dex. obtain the kinematic parameters of interest while assigning a

probability of membership PMi to each i-star by maximizing the 4. Membership and kinematic analysis following log-likelihood equation: ln L = Σ P ln [P P ] + Σ (1 − P )ln [P (1 − P )] , (1) Before proceeding with the analysis of Tucana’s kinematic and i Mi mem rad i Mi non rad chemical properties, we needed to identify the stars that are prob- where Pmem is the target’s probability distribution depending on able members of Tucana and weed out possible contaminants its l.o.s. velocity, Prad and Pnon are the prior probabilities related (foreground MW stars and background galaxies). We followed to the surface density profile of the galaxy and the presence of the same steps for the membership selection in all the catalogs possible contaminants, respectively, while PMi is defined as: we analyzed. Pmem Prad We first selected targets located approximately along the PMi = · (2) RGB of Tucana, making our selection by magnitude and color. Pmem Prad + Pnon (1 − Prad) We used a set of isochrones (Girardi et al. 2000; Bressan et al. Both equations were adapted from Eqs. (3) and (4) of Walker et al. 2012) with age tage = 12.6 Gyr and [Fe/H] ∼ −2.3 dex, and age (2009b), respectively.

A152, page 7 of 24 A&A 635, A152 (2020)

Table 3. Parameters and evidences resulting from the probability-weighted Bayesian kinematic analysis for all the datasets analyzed in this work.

Sample Nin Neff Modelv ¯hel σv k vc θ Bayes factor (km s−1) (km s−1) (km s−1 arcmin−1) (km s−1) (deg) +1.4 +1.9 − +2.4 +40 Linear 178.9−1.5 6.2−1.7 1.9−5.1 5−38 ln Blin,flat = 0.4 +1.4 +2.0 +2.3 +50 − P91 43 39 Flat 179.0−1.3 6.0−1.6 0.5−2.1 164−64 ln Brot,disp = 2.6 +1.3 +1.7 No rotation 179.0−1.3 5.7−1.5 +1.4 +2.3 +5.9 +23 Linear 180.0−1.4 7.4−1.7 3.2−2.9 173−38 ln Blin,flat = 0.6 +1.3 +1.7 +2.2 +48 − P91 + P69 63 53 Flat 180.0−1.4 6.6−1.3 1.0−2.0 152−62 ln Brot,disp = 2.3 +1.3 +1.6 No rotation 180.0−1.2 6.4−1.2 +1.5 +2.2 +4.5 +16 Linear 180.2−1.5 8.4−2.0 6.3−4.5 167−26 ln Blin,flat = 1.6 +1.3 +1.9 +2.2 +47 All comb. 109 55 Flat 179.9−1.4 7.0−1.4 1.5−2.2 140−61 ln Brot,disp = −1.1 +1.2 +1.6 No rotation 180.1−1.3 6.6−1.2 Notes. The reported values of the kinematic parameters represent the median of the corresponding marginalized posterior distributions, with 1-σ errors set as the confidence intervals around the central value enclosing 68% of each distributions.

We run the Bayesian analysis using the MultiNest code (Feroz of the kinematic major axis (i.e., the direction of the velocity et al. 2009; Buchner et al. 2014), a multimodal nested sampling gradient, perpendicular to the axis of rotation), while vrot(Ri) is algorithm, in order to obtain the kinematic parameters and the the modeled rotational velocity term. We fit and compared three membership probabilities all at once similarly to Cicuéndez et al. kinematic models: a dispersion-only model (i.e., with the veloc- (2018). A further output of this code is the Bayesian evidence, ity rotation term set to zero), a model with its rotational velocity which gives us the possibility to compare different kinematic linearly increasing with its radius (vrot(Ri) = k Ri) and a flat one models according to their statistical significance. (vrot(Ri) = vc = constant). The spatial prior probability as a function of radius Prad(Ri) The free kinematic parameters were as follows: the systemic accounts for the fact that we are more likely to observe a mem- velocityv ¯hel and velocity dispersion σv common to the three ber star near the galaxy’s center than in its outer regions. To this models, the position angle θ of the kinematic major axis, the aim, we assumed an exponentially decreasing surface-number velocity gradient k of the linear rotation model, and the con- density profile, which takes into account a uniform background stant rotational velocity vc of the flat model. In our definition, ◦ ◦ surface-number density. The parameters of the profile were the position angle θ varies between 0 and 180 , which means obtained from a VLT/VIMOS photometric dataset centered on that a rotation signal (either expressed as k or vc) that has a neg- Tucana, which was kindly provided by G. Beccari (ESO) and M. ative sign implies a receding velocity on the west side of the Bellazzini (INAF-OAS). This photometry was preferred to the galaxy (and would be equivalent to a positive gradient adding ◦ FORS2 pre-imaging and the DES-DR1 photometry used for the 180 to θ). The model evidences, Z, were combined through CMD-selection since it is much deeper (almost 3 magnitudes in the Bayes factor, B1,2 = Z1/Z2, where the subscripts indicate a I-band), although it extends up to the tidal radius. To obtain the given model. To quantify the statistical significance of one model best-fitting structural parameters we applied a Bayesian Markov with respect to another, we made use of the Jeffreys scale, based chain Monte Carlo (MCMC) analysis following the density pro- on the natural logarithm of the Bayes factor: positive values of file of the RGB stars, as was done in Cicuéndez et al.(2018), (0−1), (1−2.5), (2.5−5), (5+) correspond to inconclusive, weak, while accounting for contamination. The assumed profile proved moderate, and strong evidence favoring one model over the other to be a good representation of the observed surface-number den- (T18; Hermosa Muñoz et al. 2020, but also Wheeler et al. 2017). sity profile for Tucana, and the best-fitting structural parame- In our case, we had the following Bayes factors: ln Blin,flat, com- ters (see Table B.1) are perfectly compatible with the parameters paring the evidences of the two rotational models, and ln Brot,disp, reported by Saviane et al.(1996) 6. between the evidences of the best rotational model (choosing the The prior probability of contamination by foreground stars one that has the largest Z) and that of the dispersion-only one. P was based on the Besançon model (Robin et al. 2003). We used the following priors for the kinematic parameters: non −1 The generated distribution of l.o.s. velocities was well-fit by a −50 < (¯vhel − vg) [km s ] < 50, where vg is the initial mean −1 −1 −1 Gaussian profile (¯vBes = 57 km s ; σBes = 99 km s ). The con- value of the velocity distribution, 0 < σv [km s ] < 50, −50 < −1 −1 −1 tamination model was generated in the direction of Tucana over k [km s arcmin ] < 50, and −50 < vc [km s ] < 50. The an area equivalent to a FLAMES/GIRAFFE pointing selecting prior over θ was set iteratively: we initially chose the prior range stars over the range of colors and magnitudes described above. 0 < θ < π, ran the MultiNest code the first time in order to The l.o.s. velocity distribution of the probable member stars obtain the maximum value θm from the θ posterior distribution, Pmem(vi) was assumed to be Gaussian, as in T18, and accounted and ran the MultiNest code again while updating the prior range for the different kinematic models according to the following to −π/2 < θ − θm < +π/2. This choice accounted for the limit rotational term: vrot(Ri) cos(θ − θi), with Ri being the angular dis- case of θ near 0 or π. Results of the recovered probability-weighted kinematic tance from the galaxy’s center, θi the position angle (measured from north to east) of the ith target star, θ the position angle parameters and evidences for the three analyzed datasets, together with the effective number of probable members (defined 6 ≈ These authors also performed an exponential fit to the RGB den- as Neff ΣiPMi ) are reported in Table3. The PMi finally assigned sity profile, which, however, was obtained from a shallower photometry to each target were those obtained from the most significant kine- covering approximately the same area as that of VLT/VIMOS. matic model.

A152, page 8 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph

Fig. 3. Line-of-sight velocity distributions of the stars from the combined FORS2 + FLAMES dataset, having membership probabilities P > 0.95 (in black) and 0.05 < P < 0.95 (in red). The red solid line indicates the systemic velocity obtained from the kinematic analysis. Left panel: distribution along the optical major axis; right panel: along the minor axis.

The P91 FORS2 dataset yields 39 effective members, while when analyzing the F09 catalog. Finally, there seems to be lit- the inclusion of the P69 and FLAMES data adds about 15 more tle agreement about the systemic velocity of Tucana among F09 effective members. These numbers already double the member and G19, showing an offset of ∼20 km s−1 at more than 3-σ stars reported by F09 (17) and G19 (36, although when accounting significance. for their probability of membership they reduce to ∼20 effective As shown in Table3, our results are remarkably similar for members). We shall note that for each analyzed dataset, the effec- all the cases we analyzed. When using the P91 FORS2 data tive number of members resulted to be approximately the same alone, combining P91 and P69, or with all the datasets together, for each kinematic model. In all cases, the systemic velocity is the systemic velocity and the velocity dispersion settled around stable around 180 km s−1, while the velocity dispersion converges ∼180 km s−1 and 6 km s−1, respectively. Furthermore, we found to 6 km s−1 for the dispersion-only model, which proved to be the no significant evidence of rotation, with the dispersion-only most statistically significant one. Both the systemic velocity and model moderately favored in all of the cases. We note a slight velocity dispersion values are significantly lower than what has increase in the velocity dispersion for the rotational models: this been reported in literature studies. We refer to Sect.5 for the full is caused by few targets acquiring a higher membership prob- discussion of the results from the kinematic analysis. ability, due to the different fitted models. However, the effect We should also note that the adopted method does not take is small, and all the velocity dispersion values we obtained are into account the different selection functions of the various compatible within 1-σ. datasets that were independently built. Therefore, to test whether We further performed a simpler analysis on the three this introduced a bias in our results, we repeated the analysis, datasets, considering only those targets with the highest mem- relaxing the assumption of an exponentially declining surface- bership probabilities (i.e., with PM > 0.95) looking for the number density profile for Tucana’s stars, and simply required it kinematic parameters of a dispersion-only model, whose results to be monotonically decreasing (like in Walker et al. 2009b). No confirmed those of the probability-weighted analysis. Perform- significant difference in the results was found. ing the same test by adding those stars with a lower membership probability (i.e., with 0.05 < PM < 0.95), would instead increase −1 5. Kinematic results the velocity dispersion up to 8 km s , which is still within 1-σ of the previous results taking into account the error bars. We also The analyses of the properties of the stellar component of Tucana note that these stars are found further away from the center of carried out in the literature indicated a system with a rela- Tucana compared to the more probable members (see Fig.3), −1 tively high l.o.s. velocity dispersion (σv ∼ 15 km s ) compared but they are still inside the tidal radius of the galaxy. However, to other similarly luminous companions (see e.g., the compila- it is hard to discern whether the increase in the velocity disper- tion for LG dwarfs of Kirby et al. 2014; Wheeler et al. 2017). The sion seen when including stars with lower membership probabil- works of F09 and G19 also reported the tentative presence of a ity is caused by a radially increasing velocity-dispersion profile, velocity gradient likely due to internal rotation, since perspective or simply because (all or part of) these stars are contaminants. effects related to transverse motion are negligible at the distance Considering those stars with a lower membership probability of Tucana. However, in a Bayesian analysis of the rotational sup- that have metallicity measurements, they seem to be preferen- port of the stellar component of LG galaxies using data from the tially metal-poor (see Fig.4). Therefore, it may also be that the literature, Wheeler et al.(2017) found no significant evidences increase in σv could be caused by the preferential inclusion of for rotation and a l.o.s. velocity dispersion as high as 21 km s−1 metal-poor stars with a hotter velocity dispersion than the more

A152, page 9 of 24 A&A 635, A152 (2020)

semi-minor axis). We chose an underlying linear rotation com- ponent since it resulted in higher evidence compared to a flat rotation model when analyzing our data. Each case was simu- lated N = 1000 times, in which we run our Bayesian kinematic analysis fitting just the linear rotation and the dispersion-only models in order to recover the related parameters and evidences. Results are shown in Table B.2. Results from the tests indicate that the linear rotation would be spotted with high significance for velocity gradient values ≥5.6 km s−1 arcmin−1 aligned with the projected optical major axis. If the underlying rotation is instead milder, for exam- ple with velocity gradients ≤3.7 km s−1 arcmin−1, the recovered evidences for rotation are weak, inconclusive, or favoring the dispersion-only model, as we move through decreasing values of k and through different position angles. In any case, it is evident that if Tucana has a weak rotation signal (i.e., k ≤ 3.7 km s−1 arcmin−1), with the data we had, we could not detect Fig. 4. Metallicity versus heliocentric velocity for the observed target it with high significance, particularly if it is not aligned with of the combined FORS2 dataset. Black dots represent the targets with the optical major axis. If instead Tucana has a velocity gradi- P > 0.95, while red dots show those with 0.05 < P < 0.95. ent as reported in F09 and G19, meaning with a value close to k = 5.6 km s−1 arcmin−1 along the major axis, with our data we would have detected it with high significance, but this was not metal-rich stars (Tolstoy et al. 2004; Battaglia et al. 2006, 2008, the case. Therefore, it seems that if rotation is actually present in 2011; Amorisco & Evans 2012a). We refer to Sect. 6.2 for more Tucana, it is probably at a level of vrot/σv . 0.5, and to detect it details on this point. with high significance a better sampling would be needed, in par- We highlight the fact that underestimated (optimistic) or ticular at radii around 3 . R/Re . 5. We note that these results overestimated (pessimistic) velocity errors may impact on the are conservative: in fact, if we had considered an input velocity measured velocity dispersion. We refer the reader to Fig. 1 in dispersion as high as 10 km s−1, the linear rotation signal would Koposov et al.(2011) for an analysis of the impact of under- have been even more difficult to detect. estimated/overestimated velocity errors as a function of the ratio between the true error and the true velocity dispersion. However, as reported in AppendixA, we conducted several consistency 5.2. Comparison with other works tests where we show that the velocity errors are well-determined. In Appendix A.3, we perform a comparative analysis of the Therefore, our results for Tucana point to a value for the velocity measurements for the individual stars derived in this −1 velocity dispersion which is very unlikely to exceed 10 km s . work and those in F09 and G19. Here, we focus instead on the +1.6 −1 We assume σv = 6.2−1.3 km s as our reference value, averaging comparison of the recovered velocity parameters from the kine- between the results of the dispersion-only model from the ana- matic analyses of the different works. lyzed datasets. Firstly, we ran our code on the F09 velocity measurements for member stars, and the same for G19, in order to see if we 5.1. MultiNest mock tests were able to recover their results. Considering a dispersion- only model, we found a 1-σ agreement between the recovered Although we have found that the rotation signal in our catalog is velocity parameters and those reported by these authors. There- not statistically significant, we conducted a series of mock tests fore, our procedure does not introduce a bias, and we can directly in order to explore which rotational properties can be detected compare with the reported values in F09 and G19. according to the characteristics of our data. To this aim, we We found an offset of ∼15 (35) km s−1 for the systemic veloc- followed the approach already introduced in T18 and Hermosa ity between the values reported by F09 (G19) and by us – −1 +2.9 −1 Muñoz et al.(2020): we produced mock catalogs of l.o.s. veloc- v¯hel,F09 = 194.0 ± 4.3 km s andv ¯hel,G19 = 216.7−2.8 km s . ities assuming the same number, spatial position, velocity dis- These offsets are somewhat higher but still compatible with those tribution, and velocity uncertainties of the observed data. Our reported in Appendix A.3, so we refer the reader to that section base catalog was the combined FORS2 P91 + P69 + FLAMES for an analysis of the possible causes. We stress that if we ana- dataset after applying a PM > 0.05 cut; the inclusion of less prob- lyze our reduction of the P69 and the FLAMES data on their able members was a compromise to have the highest number own, we obtain systemic velocities compatible at the 1-σ level of targets (57) within the largest spatial area. To each target we with our value of ∼180 km s−1; therefore, the differences encoun- assigned a mock velocity vmock randomly extracted from a Gaus- tered appear to be related to the treatment of the datasets. On the sian distribution centered on zero and of standard deviation equal other hand, the velocity dispersion values (without accounting −1 to the assumed velocity dispersion, fixed at σv,mock = 6 km s . for the presence of possible gradients) reported by F09 and G19 +4.1 −1 +2.8 −1 This value of σv,mock was set according to the converging results – σv,F09 = 15.8−3.1 km s and σv,G19 = 14.4−2.3 km s , differ from the probability-weighted analysis of our datasets. We fur- from our reference σv value almost at the 3-σ level. ther added a projected linear rotational component vrot,mock, such If we were to analyze our reduction of the P69 FORS2 that vrot,mock/σv,mock = n = 1.5, 1.0, 0.75, 0.5, 0.25, 0 at the half- dataset alone, we would obtain 17 effective members out of 23 light radius. These correspond to velocity gradients of k = input targets, showing a velocity dispersion value of σv,P69 = −1 −1 +3.7 −1 11.2, 7.5, 5.6, 3.7, 1.9, 0 km s arcmin , which we simulated at 11.1−2.7 km s , associated with a highly significant rotation three different position angles, starting from the PA of the opti- signal. We noticed, however, that this gradient is driven by just cal semi-major axis (97◦) and then adding 45 and 90◦ (optical two targets that have a very low membership probability when

A152, page 10 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph analyzing the combined FORS2 dataset. If we excluded them, the 102 velocity dispersion would drop to ∼8 km s−1 and the velocity gra- dient basically disappears. Therefore, it seems here that the low- number statistics strongly limit the conclusions we could get on the kinematic status of Tucana via the P69 dataset on its own. ]

The comparison with G19 results are even more puzzling. s /

If we were to apply the same exercises as for the P69 dataset, m 101 k [ analyzing the FLAMES data on their own, we would obtain 11 c effective members out of 58 targets, which, however, could not V resolve the σv value. This is probably due to the combination of a small number of effective members and the fact that the aver- Local Group (Kirby et al. 2014) age velocity error of stars with a high probability of membership Cetus (Taibi et al. 2018) −1 (δv ∼ 6 km s ) is comparable to the σv value we find in our Tucana (Taibi et al. 2020) 100 main kinematic analysis. 102 103 Furthermore, we want to underline that we found several dif- R1/2 [pc] ferences when comparing our velocity measurements to those of G19, as described in Appendix A.3. We suspect that the way the data were actually sky subtracted has led to an excess of stars 30 40 50 60 70 80 with velocities around 220 km s−1 in the G19 work, probably due Vmax [km/s] to a combination of sky-line residuals and low S/N, which would Fig. 5. Circular velocity at half light radius (under assumption of flat σ have created fake CaT features. profile, isotropy, and sphericity) for Local Group galaxies from Kirby et al.(2014) and the isolated dwarfs Cetus and Tucana, from Taibi et al. 5.3. Implications for Tucana’s dark matter halo properties (2018) and this work. Circular velocity profiles for NFW halos between 8.5 ≤ log(Mvir/M ) ≤ 11 and following a mass concentration relation As previously discussed, the analysis of the internal kinematic from Dutton & Macciò(2014) are shown color-coded by Vmax. Local properties of Tucana yields consistent results across the com- Group galaxies represented as gray filled circles. Cetus and Tucana are bination of datasets. Our best value for the velocity dispersion highlighted with a black triangle and square, respectively; they occupy +1.6 −1 locations comparable to other dwarfs of similar stellar mass (see also of σv = 6.2−1.3 km s is significantly lower (∼3-σ) than what was reported in the literature by both F09 and G19 (see also Fig. B.2). the discussion in the previous section), but closer now to the values observed for other similarly luminous dwarf galaxies of the Local Group (see e.g. Kirby et al. 2014; Revaz & Jablonka Navarro-Frank-White halos sampled from the halo mass concen- 2018). tration relation of Dutton & Macciò(2014) are also shown (see V We use the Wolf et al.(2010) mass-estimator, valid for Appendix B.2 for sampling details) and color-coded by max. dispersion-supported spherical systems, to calculate Tucana’s The updated velocity dispersion measurements for Tucana and Cetus place these galaxies in a locus occupied by comparable dynamical mass inside the half-light radius, M1/2 = −1 2 −1 2 luminosity dwarf galaxies. These results would imply not only 3G σvr1/2 ≈ 4G σv Re, where G is the gravitational con- stant and r is the 3D de-projected half-light radius, which that Tucana does not reside in a very centrally dense halo as pre- 1/2 dicted by G19, but also that this galaxy it is not an exception to can be approximated to 4/3 Re. Using the values from Table1 +0.4 7 the too-big-to-fail problem (see e.g., Kirby et al. 2014) and substituting for σv, we obtain M1/2 = 0.7−0.3 × 10 M , which corresponds to a mass-to-light ratio within the half-light Cetus and Tucana’s isolation and well measured SFHs offer +8 an intriguing leverage to potentially separate the multitude of radius of M1/2/LV = 13−7 M /L , assuming a luminosity of 5 solutions to the too-big-to-fail problem. In particular, what can LV = 5.5 ± 1.5 × 10 L (adapted from Saviane et al. 1996). be expected from any environmental or SFH dependence on Using instead the velocity dispersion values from F09 and the galaxy density profile in baryonic feedback scenarios (see G19, applying the Wolf et al.(2010) mass-estimator we would discussion in Read et al. 2019), and how this could contrast M . +2.6 × 7 M M . +1.6 × obtain 1/2,F09 = 4 8−2.0 10 and 1/2,G19 = 3 9−1.4 with self-interacting dark matter solutions, which could predict 7 10 M respectively, which are more than four time as large as more homogeneous behavior among all dwarfs. Further assess- our own best estimation of Tucana’s mass. ing these scenarios in light of the results for the isolated dSphs, Our measurements for velocity dispersion and dynamical and their spatial stellar population distributions will be the focus mass of Tucana provide a new perspective to the discussion of a follow-up project. found in the literature. If instead of√ M1/2 we use the value of the circular velocity Vcirc(r1/2) = 3σv, we obtain a value of +3 −1 6. Chemical analysis 11−2 km s for Tucana, assuming our reference value for σv. This Vcirc(r1/2) is comparable to those of other similarly lumi- The analysis of the [Fe/H] values of the FORS2 combined dataset nous dwarfs like Carina, Sextans, and Leo II, as it is possible to led to the following results: considering the values with PM > see from Fig. 10 in G19, but also from Fig. 1 in Boylan-Kolchin 0.95, we obtained median [Fe/H] = −1.58 dex, σMAD = 0.47 dex, et al.(2012). From this last reference, we find that Tucana σintrinsic = 0.39 dex; while adding the stars with 0.05 < P < 0.95 should live in a NFW dark matter halo with a maximum circu- we instead obtained a median of [Fe/H] = −1.61 dex, σMAD = −1 lar velocity Vmax ≤ 24 km s , as is the case for many dSphs of 0.48 dex, σintrinsic = 0.39 dex. Therefore, Tucana is a metal-poor the LG. system with a significant spread in metallicity. The median [Fe/H] In Fig.5, we show the position of Tucana on the Vcirc − r1/2 value measured for the likely members is in very good agree- plane with respect to the Local Group compilation of Kirby et al. ment with the integrated quantity derived from Tucana’s SFH: (2014) and the updated value for Cetus from Taibi et al.(2018). h[Fe/H]i = −1.52 ± 0.07 dex (Monelli et al. 2010). In addition,

A152, page 11 of 24 A&A 635, A152 (2020) our average [Fe/H] value falls within the rms scatter of the stellar GPR Gaussian kernel luminosity-metallicity relation for LG dwarf galaxies reported by GPR Gaussian kernel GPR - 1 C.I. Kirby et al.(2013a), while the intrinsic scatter agrees well with the 1.0 GPR - 1 C.I. values of other similarly luminous dwarf galaxies (Leaman et al. P > 0.95 2013). 0.05 < P < 0.95

The distribution of the [Fe/H] values as a function of the 1.5 elliptical radius is shown in Fig.6. The significant scatter in metallicity toward the inner part of the galaxy is evident. In addi- tion, there is a bimodality in the metallicity histogram suggest- [Fe/H] 2.0 ing the presence of two subpopulations. This can be related to the results obtained from deep photometric data (Monelli et al. 2010), where it was shown that the splitting observed in the HB, 2.5 in the RGB bump, and in the properties of the RR-Lyrae stars, imply that Tucana has been able to produce a second generation of metal-rich stars thanks to the self-enrichment from the first 3.0 stars in a very short period of star formation (∼1 Gyr). In a more 1 2 3 4 5 0 5 recent study, Savino et al.(2019) reanalyzed the photometric R/Re N data from Monelli et al.(2010), by studying the main-sequence Fig. 6. [Fe/H] as a function of the elliptical radius scaled with Re for turn off and the HB of Tucana. They were able to obtain an Tucana’s probable member stars from the FORS2 combined dataset. SFH with a finer temporal resolution, showing that Tucana actu- Black dots represent the targets with membership probabilities P > ally experienced two early phases of star formation (SF), fol- 0.95, while the red dots show those with 0.05 < P < 0.95. The black lowed by a third one ending between 6 and 8 Gyr ago, with the solid line represents the result of a Gaussian process regression analy- two initial episodes being the most intensive. According to the sis using a Gaussian kernel and taking into account an intrinsic scatter; age-metallicity relation recovered by Savino et al.(2019), our the gray band indicates the corresponding 1-σ confidence interval. The [Fe/H] measurements could be related to the intermediate-old solid red line and the red band indicate the same, but using all targets with P > 0.05. The histogram on the right side represents the metallicity and intermediate-young age populations of the two last episodes distribution of the stars with P > 0.95. of SF. This would explain the bimodality we observed in the metallicity histogram, while our most metal-poor stars could be related to the oldest episode of SF. However, despite the rela- being a kernel-based non-parametric probabilistic method that tively high intensity of this SF period, we found very few stars makes it possible to compute empirical confidence intervals. with [Fe/H] < −2.25 dex. This is probably related to the fact that, Since we are looking for a smooth function, it performs better for lower metallicity stars, due to weaker lines, a higher S/N than a least-square fit in finding the general trend in the data. is needed in order to get similar accuracy in [Fe/H] measure- In our case, we confirmed the decreasing trend, although the 1- ments. Furthermore, their spatial distribution tend to be more σ confidence limits proved quite large due to the high intrinsic extended, which requires extra attention during the sampling scatter of the data, making the presence of a metallicity gradient phase. within ∼3 Re dubious. We further checked this result by performing a simple sim- 6.1. Looking for a metallicity gradient ulation. We assumed a double Gaussian metallicity distribution with parameters roughly fitting the observed one, but no spa- Observations have shown that in many LG dwarf galaxies the tial variation (µ[Fe/H],1 = −2.0 dex, σ[Fe/H],1 = 0.2, µ[Fe/H],2 = young and metal-rich stars are more spatially concentrated than −1.3 dex, σ[Fe/H],2 = 0.2, assuming the same fraction of stars the old and metal-poor ones that display a more extended spa- in the two Gaussians). We then randomly extracted [Fe/H] val- tial distribution. Their overall radial distribution produces a ues at the radial positions of our data. We further reshuffled the decreasing metallicity gradient, e.g., Fornax (Battaglia et al. [Fe/H] values according to the observed errors, and, finally, we 2006; Leaman et al. 2013), Phoenix (Kacharov et al. 2017), performed a linear least-square fit looking for a spatial metallic- which could eventually reach a plateau on the outside, e.g., ity gradient. We repeated this process 1000 times. The obtained Sculptor (Tolstoy et al. 2004), VV 124 (Kirby et al. 2013b), average gradient was compatible with zero, with the associated Cetus (T18). These results have also been reproduced by simula- scatter large enough to include within 1-σ the observed value of tions (e.g., Schroyen et al. 2013; Revaz & Jablonka 2018), which m. Therefore, with the data we had, the observed gradient within have shown that the shape of these gradients strongly depends on R < 3 Re is not statistically significant. the combination of the stellar mass, SFH, and dynamical history Adding the stars with 0.05 < P < 0.95 would extend of the system under consideration, although their strength could the spatial coverage up to R ∼ 6 Re, thanks to the two outer- be influenced by merger events (Benítez-Llambay et al. 2016) or most targets, but would lead to an even milder gradient: m = environmental effects such as tidal stripping (Sales et al. 2010). −0.07 ± 0.04 dex arcmin−1 (= −0.28 ± 0.16 dex kpc−1 = −0.06 ± Therefore, we investigated the presence of a metallicity gra- −1 0.03 dex Re ), by performing a linear least-square fit. The pres- dient as a function of radius, first focusing on the stars with ence of a metallicity gradient in Tucana is expected from stud- P > 0.95, which extended up to ∼3 Re (see Fig.6). Perform- ies of deep-photometric data (Hidalgo et al. 2013; Savino et al. ing an error-weighted linear least-square fit to the data, we 2019). However it is probable that we are mainly targeting stars d[Fe/H] −1 obtained the value m = dR = −0.16 ± 0.09 dex arcmin belonging to the more recent episodes of SF, whose populations −1 −1 (= −0.6 ± 0.4 dex kpc = −0.13 ± 0.07 dex Re , using the val- share similar spatial extensions (see e.g., Fig. 11b in Savino ues reported in Table1 for the conversions). We also performed et al. 2019). Therefore, the presence of a metallicity gradient in a Gaussian process regression (GPR) analysis, where we used a Tucana is very tentative, and we would need a better sampling, Gaussian kernel together with a noise component to account for in particular, of the metal-poor component, around 3 . R/Re . 5 the intrinsic metallicity scatter. The GPR has the advantage of to put our results on firmer ground.

A152, page 12 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph

We compared Tucana’s metallicity gradient (or rather the Table 4. Parameters from the Bayesian kinematic analysis of the metal- lack thereof) with those of some MW satellites with similar rich (MR) and metal-poor (MP) sub-samples defined in Sect. 6.2. 5 luminosities (LV . 5 × 10 L ) and short SFHs, for example, Draco (Aparicio et al. 2001), Ursa Minor (Carrera et al. 2002), P > 0.95 P > 0.05 and Sextans (Bettinelli et al. 2018). All of them formed the majority of their stars more than 10 Gyr ago within a short period Sample N σv N σv − − of SF, which in some cases may have lasted no more than 1 Gyr (km s 1) (km s 1) (i.e., Sextans). It has been shown that such short SFHs may MR 26 6.0+1.9 27 5.9+1.9 lead to mild metallicity gradients in these systems (Marcolini −1.6 −1.6 MP 26 9.0+2.8 27 10.9+2.6 et al. 2008; Kirby et al. 2011; Revaz & Jablonka 2018). Indeed, −2.4 −2.2 both Draco and Ursa Minor show mild gradients with values of −1 −1 −0.05 dex Re and −0.03 dex Re , as reported by Schroyen et al. 7 +1.9 −1 +2.6 −1 (2013) and Kirby et al.(2011), respectively . Sextans, on the we obtained: σv,MR = 5.9 km s and σv,MP = 10.9 km s , −1 −1.6 −2.2 other hand, seems to have a stronger gradient of −0.24 dex Re , which are instead ∼2-σ from each other (see Table4). as reported by Schroyen et al.(2013) using the Battaglia et al. Therefore, there is weak evidence of two chemo- (2011) spectroscopic dataset. However, this value is somewhat kinematically distinct subpopulations in Tucana. Additional overestimated since recent studies of the structural properties of data, in particular including external parts of the galaxy are Sextans (Roderick et al. 2016; Cicuéndez et al. 2018) show that necessary to reach firm conclusions, regarding the presence of this system is less extended than what was previously reported in a metallicity gradient as well as the possible distinct chemo- the literature. Using the half-light radius value from Cicuéndez kinematic populations. −1 et al.(2018), we find a lower gradient of −0.18 dex Re , which is still far from the other dwarf’s values and probably related to an early merger event that could have steepened its metallic- ity gradient (see Cicuéndez & Battaglia 2018, but also Benítez- 7. Summary and conclusions Llambay et al. 2016). If the case of Tucana is similar to those of In this paper, we present results regarding the internal kinematic Draco and Ursa Minor, as it seems, we would expect it to host and metallicity properties of the Tucana dwarf spheroidal galaxy, an equally mild metallicity gradient, but it would take a better based on the analysis of multi-object spectroscopic samples of sampling of the spatial extension of the metal-poor component individual RGB stars. in Tucana to confirm it. This analysis is based on a novel set of 50 individual objects collected with the VLT/FORS2 instrument in MXU 6.2. Searching for two chemo-kinematically-distinct mode in P91, complemented by a re-reduction and reanalysis populations of two datasets from the literature, namely the VLT/FORS2- MXU dataset presented in Fraternali et al.(2009, F09) and the Although we have not spotted a clear metallicity gradient in VLT/FLAMES-GIRAFFE one by Gregory et al.(2019, G19). Tucana, the bimodality found in the metallicity distribution, may Applying a probabilistic membership approach, we find 39 effec- indicate the presence of two subpopulations which differ not tive members in our P91 sample, which doubles the number of only in their chemical properties, but also in their kinemat- Tucana’s member stars found in F09 and G19. ics. Some of the dSphs satellites of the MW, such as Sculp- A full re-reduction and analysis of the data presented in tor, Fornax, Carina, and Sextans, show such features where the the literature was carried out because it became clear that the metal-rich (usually more spatially concentrated) subpopulation published catalogs could not be directly combined with the has colder kinematics than the metal-poor (and more extended) l.o.s. velocities we derived for the P91 sample: there are sig- one (e.g., Tolstoy et al. 2004; Battaglia et al. 2006, 2008, 2011; nificant differences between the values of the systemic velocity Koch et al. 2008; Amorisco & Evans 2012a). Determining the reported in those studies with respect to that we derived from the chemo-kinematic properties of dSphs is of great interest, not P91 dataset (∼195 km s−1 for F09, ∼215 km s−1 for G19, while only to better understand their evolutionary path, but also to get ∼180 km s−1 in our case); and the comparison of the individual an insight into their dark matter properties (e.g., Battaglia et al. l.o.s. velocities for the stars in common were supporting the pres- 2008; Walker & Peñarrubia 2011; Amorisco & Evans 2012b; ence of shifts with respect to F09, but were not sufficient to fully Strigari et al. 2018). quantify whether that was the only source of difference, or if it In the case of Tucana, we first analyzed the combined FORS2 was even more unfavorable for the comparison with G19. dataset with the P > 0.95 cut applied (see Fig.4). We took Following our homogeneous data reduction, we find an the median [Fe/H] value of −1.58 dex to split our sample into excellent agreement between velocity measurements of the three metal-rich (MR) and a metal-poor (MP) subsamples. We then datasets both for stars in common (Fig. A.1 top row panels) as ran our code to obtain the kinematic parameters of both sam- well as for systemic velocity, which is stable around 180 km s−1 ples (see Sect.5). Using the dispersion-only model, we found for the three datasets (see Fig.2, right panel, and Table3). We +1.9 −1 +2.8 −1 σv,MR = 6.0−1.6 km s and σv,MP = 9.0−2.4 km s , which are 1- proceeded to analyze the P91 dataset alone and also in combi- σ from each other. Including instead the 0.05 < P < 0.95 data, nation with our treatment of the P69 and FLAMES data. The resulting values of the intrinsic l.o.s. velocity dispersion are −1 7 consistently around 6 km s when considering the three com- We revised Draco’s gradient using the more recent and spatially binations of datasets (P91, P91 + P69, all the three combined) extended dataset of Walker et al.(2015): we performed a broad mem- bership selection like in Walker et al.(2015; see their Fig. 10) and and the highly probable members (probability of membership then refined it by cross-correlating with the Gaia-DR2 catalog, select- PM > 0.95); when including lower probability members, the ing those targets co-moving with Draco. We found a value of −0.09 ± velocity dispersion increases, but it is unlikely that σv is higher −1 −1 0.02 dex Re from the linear least-square fit, in fair agreement with the than 10 km s . Therefore, our analysis leads to the conclusion previous one. that the l.o.s. velocity dispersion of Tucana’s stellar component

A152, page 13 of 24 A&A 635, A152 (2020) is much lower than the values reported by F09 and G19 – References σ . +4.1 −1 σ . +2.8 −1 v,F09 = 15 8−3.1 km s and v,G19 = 14 4−2.3 km s . Abbott, T. M. C., Abdalla, F. B., Allam, S., et al. 2018, ApJS, 239, 18 Furthermore, we find no significant signs of internal rota- Amorisco, N. C., & Evans, N. W. 2012a, ApJ, 756, L2 tion. Mock tests suggest that if Tucana had had a maximum rota- Amorisco, N. C., & Evans, N. W. 2012b, MNRAS, 419, 184 tional velocity of ∼10−15 km s−1 along the projected major axis Aparicio, A., Carrera, R., & Martínez-Delgado, D. 2001, AJ, 122, 2524 (as previously reported in literature) with the data available to Battaglia, G., Tolstoy, E., Helmi, A., et al. 2006, A&A, 459, 423 Battaglia, G., Helmi, A., Tolstoy, E., et al. 2008, ApJ, 681, L13 us, we should have detected it with high significance. On the Battaglia, G., Tolstoy, E., Helmi, A., et al. 2011, MNRAS, 411, 1013 other hand, lower levels of rotation are not completely ruled out, Battaglia, G., Helmi, A., & Breddels, M. 2013, New Astron. Rev., 57, 52 but a larger sample would be needed to quantify their presence. Battaglia, G., Sollima, A., & Nipoti, C. 2015, MNRAS, 454, 2401 Nevertheless, it seems improbable that Tucana is a fast rotator Benítez-Llambay, A., Navarro, J. F., Abadi, M. G., et al. 2016, MNRAS, 456, 1185 (vrot/σv & 1). +1.6 −1 Bermejo-Climent, J. R., Battaglia, G., Gallart, C., et al. 2018, MNRAS, 479, Assuming an average σv = 6.2−1.3 km s for our data, we 1514 obtain a dynamical mass within the half-light radius of M1/2 = Bernard, E. J., Monelli, M., Gallart, C., et al. 2009, ApJ, 699, 1742 +0.4 7 Bettinelli, M., Hidalgo, S. L., Cassisi, S., Aparicio, A., & Piotto, G. 2018, 0.7−0.3 × 10 M . This translates into a circular velocity at the half-light radius of V (r ) = 11+3 km s−1, which implies that, MNRAS, 476, 71 circ 1/2 −2 Boylan-Kolchin, M., Bullock, J. S., & Kaplinghat, M. 2012, MNRAS, 422, if Tucana inhabits a NFW dark matter halo, it should have a sim- 1203 ilar density to those of other MW dSphs (Boylan-Kolchin et al. Bressan, A., Marigo, P., Girardi, L., et al. 2012, MNRAS, 427, 127 2012). Therefore, Tucana is not an exception to the too-big-to- Brooks, A. M., & Zolotov, A. 2014, ApJ, 786, 87 fail problem and not “a massive failure”, as it had gained fame Buchner, J., Georgakakis, A., Nandra, K., et al. 2014, A&A, 564, A125 Carrera, R., Aparicio, A., Martínez-Delgado, D., & Alonso-García, J. 2002, AJ, for being. 123, 3199 The analysis of Tucana’s chemical properties was only car- Cicuéndez, L., & Battaglia, G. 2018, MNRAS, 480, 251 ried out on the combined P91 and P69 FORS2 data, due to Cicuéndez, L., Battaglia, G., Irwin, M., et al. 2018, A&A, 609, A53 their higher S/N. We establish that the galaxy is mainly metal- Dutton, A. A., & Macciò, A. V. 2014, MNRAS, 441, 3359 poor with a significant scatter in metallicity (having a median Feroz, F., Hobson, M. P., & Bridges, M. 2009, MNRAS, 398, 1601 Fraternali, F., Tolstoy, E., Irwin, M. J., & Cole, A. A. 2009, A&A, 499, 121 [Fe/H] = −1.58 dex, σMAD = 0.47 dex and σintrinsic = 0.39 dex Fritz, T. K., Battaglia, G., Pawlowski, M. S., et al. 2018, A&A, 619, A103 when considering only highly likely members, and median Gaia Collaboration (Helmi, A., et al.) 2018, A&A, 616, A12 [Fe/H] = −1.61 dex, σMAD = 0.48 dex, and σintrinsic = 0.39 dex Gallart, C., Monelli, M., Mayer, L., et al. 2015, ApJ, 811, L18 when including the less likely members). The derived values Garrison-Kimmel, S., Boylan-Kolchin, M., Bullock, J. S., & Lee, K. 2014, agree very well with SFH studies (Monelli et al. 2010; Savino MNRAS, 438, 2578 Girardi, L., Bressan, A., Bertelli, G., & Chiosi, C. 2000, A&AS, 141, 371 et al. 2019). In addition, the average [Fe/H] falls between the Gregory, A. L., Collins, M. L. M., Read, J. I., et al. 2019, MNRAS, 485, rms scatter of the stellar luminosity-metallicity relation for LG 2010 dwarf galaxies (Kirby et al. 2013a). Hanuschik, R. W. 2003, A&A, 407, 1157 Looking at the distribution of the [Fe/H] values as a func- Hausammann, L., Revaz, Y., & Jablonka, P. 2019, A&A, 624, A11 Hendricks, B., Koch, A., Walker, M., et al. 2014, A&A, 572, A82 tion of radius, we find a mild metallicity gradient. However, Hermosa Muñoz, L., Taibi, S., Battaglia, G., et al. 2020, A&A, 634, A10 the size and spatial distribution of the current datasets do not Hidalgo, S. L., Monelli, M., Aparicio, A., et al. 2013, ApJ, 778, 103 lead to a statistically significant detection. The presence of a Holtzman, J. A., Afonso, C., & Dolphin, A. 2006, ApJS, 166, 534 gradient in Tucana would be expected from the age gradients Iorio, G., Nipoti, C., Battaglia, G., & Sollima, A. 2019, MNRAS, 487, inferred from deep photometric studies (Hidalgo et al. 2013; 5692 Jordi, K., Grebel, E. K., & Ammon, K. 2005, Astron. Nachr., 326, 657 Savino et al. 2019), but also from both observations and sim- Kacharov, N., Battaglia, G., Rejkuba, M., et al. 2017, MNRAS, 466, 2006 ulations of similarly luminous dwarfs (see e.g., Leaman et al. Kazantzidis, S., Łokas, E. L., Callegari, S., Mayer, L., & Moustakas, L. A. 2011, 2013; Schroyen et al. 2013; Revaz & Jablonka 2018), which ApJ, 726, 98 indeed host mild metallicity gradients. Therefore, the presence Kazantzidis, S., Mayer, L., Callegari, S., Dotti, M., & Moustakas, L. A. 2017, ApJ, 836, L13 of an underling gradient in Tucana is not excluded, but it would Kirby, E. N., Lanfranchi, G. A., Simon, J. D., Cohen, J. G., & Guhathakurta, P. need a better sampling of the metal-poor component in Tucana 2011, ApJ, 727, 78 (particularly at R & 3 Re) to confirm it. Finally, we find a Kirby, E. N., Cohen, J. G., Guhathakurta, P., et al. 2013a, ApJ, 779, 102 hint of the presence of multiple stellar populations with distinct Kirby, E. N., Cohen, J. G., & Bellazzini, M. 2013b, ApJ, 768, 96 chemo-kinematical properties, although in this case, the addi- Kirby, E. N., Bullock, J. S., Boylan-Kolchin, M., Kaplinghat, M., & Cohen, J. G. 2014, MNRAS, 439, 1015 tion of new data would also help to put this result on firmer Klimentowski, J., Łokas, E. L., Kazantzidis, S., Mayer, L., & Mamon, G. A. ground. 2009, MNRAS, 397, 2015 Koch, A., Grebel, E. K., Gilmore, G. F., et al. 2008, AJ, 135, 1580 Koposov, S. E., Gilmore, G., Walker, M. G., et al. 2011, ApJ, 736, 146 Acknowledgements. We would like to thank the anonymous referee for the Lavery, R. J., & Mighell, K. J. 1992, AJ, 103, 81 helpful and constructive comments. We thank Giacomo Beccari and Michele Leaman, R., Venn, K. A., Brooks, A. M., et al. 2012, ApJ, 750, 33 Bellazzini for sharing with us their VLT/VIMOS photometric data. We also thank Leaman, R., Venn, K. A., Brooks, A. M., et al. 2013, ApJ, 767, 131 F. Fraternali, E. Tolstoy, A. Gregory and M. Collins for useful discussions in the Lemasle, B., Hill, V., Tolstoy, E., et al. 2012, A&A, 538, A100 different phases of this project. S.T. thankfully acknowledges ESO for a one Lemasle, B., de Boer, T. J. L., Hill, V., et al. 2014, A&A, 572, A88 month visit funded by an SSDF grant. G.B. and S.T. acknowledge financial sup- Lewis, G. F., Ibata, R. A., Chapman, S. C., et al. 2007, MNRAS, 375, 1364 port through the grants (AEI/FEDER, UE) AYA2017-89076-P, AYA2014-56795- Lux, H., Read, J. I., & Lake, G. 2010, MNRAS, 406, 2312 P, as well as by the Ministerio de Ciencia, Innovación y Universidades (MCIU), Marcolini, A., D’Ercole, A., Battaglia, G., & Gibson, B. K. 2008, MNRAS, 386, through the State Budget and by the Consejería de Economía, Industria, Com- 2173 ercio y Conocimiento of the Canary Islands Autonomous Community, through Mateo, M. L. 1998, ARA&A, 36, 435 the Regional Budget. S.T. acknowledges an FPI fellowship BES-2015-074765, Mayer, L., Governato, F., Colpi, M., et al. 2001a, ApJ, 559, 754 while G.B. acknowledges financial support through the grant RYC-2012-11537. Mayer, L., Governato, F., Colpi, M., et al. 2001b, ApJ, 547, L123 This research has made use of NASA’s Astrophysics Data System and extensive Mayer, L., Mastropietro, C., Wadsley, J., Stadel, J., & Moore, B. 2006, MNRAS, use of IRAF, Python, Numpy, Scipy and Astropy ecosystems. 369, 1021

A152, page 14 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph

McConnachie, A. W. 2012, AJ, 144, 4 Sawala, T., Frenk, C. S., Fattahi, A., et al. 2016, MNRAS, 457, 1931 Monelli, M., Gallart, C., Hidalgo, S. L., et al. 2010, ApJ, 722, 1864 Schroyen, J., De Rijcke, S., Koleva, M., Cloet-Osselaer, A., & Vandenbroucke, Muñoz, R. R., Majewski, S. R., & Johnston, K. V. 2008, ApJ, 679, 346 B. 2013, MNRAS, 434, 888 Noll, S., Kausch, W., Kimeswenger, S., et al. 2014, A&A, 567, A25 Simon, J. D. 2018, ApJ, 863, 89 Pasetto, S., Grebel, E. K., Berczik, P., Chiosi, C., & Spurzem, R. 2011, A&A, Spencer, M. E., Mateo, M., Walker, M. G., & Olszewski, E. W. 2017, ApJ, 836, 525, A99 202 Piatek, S., & Pryor, C. 1995, AJ, 109, 1071 Starkenburg, E., Hill, V., Tolstoy, E., et al. 2010, A&A, 513, A34 Read, J. I., & Erkal, D. 2019, MNRAS, 487, 5799 Strigari, L. E., Frenk, C. S., & White, S. D. M. 2018, ApJ, 860, 56 Read, J. I., Wilkinson, M. I., Evans, N. W., Gilmore, G., & Kleyna, J. T. 2006, Taibi, S., Battaglia, G., Kacharov, N., et al. 2018, A&A, 618, A122 MNRAS, 367, 387 Teyssier, M., Johnston, K. V., & Kuhlen, M. 2012, MNRAS, 426, 1808 Read, J. I., Walker, M. G., & Steger, P. 2019, MNRAS, 484, 1401 Tolstoy, E., Irwin, M. J., Helmi, A., et al. 2004, ApJ, 617, L119 Revaz, Y., & Jablonka, P. 2018, A&A, 616, A96 Tolstoy, E., Hill, V., & Tosi, M. 2009, ARA&A, 47, 371 Revaz, Y., Jablonka, P., Sawala, T., et al. 2009, A&A, 501, 189 Walker, M. G. 2013, in Dark Matter in the Galactic Dwarf Spheroidal Satellites, Robin, A. C., Reylé, C., Derrière, S., & Picaud, S. 2003, A&A, 409, 523 eds. T. D. Oswalt, & G. Gilmore, 5, 1039 Rocha, M., Peter, A. H. G., & Bullock, J. 2012, MNRAS, 425, 231 Walker, M. G., & Peñarrubia, J. 2011, ApJ, 742, 20 Roderick, T. A., Jerjen, H., Da Costa, G. S., & Mackey, A. D. 2016, MNRAS, Walker, M. G., Mateo, M., & Olszewski, E. W. 2008, ApJ, 688, L75 460, 30 Walker, M. G., Mateo, M., & Olszewski, E. W. 2009a, AJ, 137, 3100 Sales, L. V., Navarro, J. F., Abadi, M. G., & Steinmetz, M. 2007, MNRAS, 379, Walker, M. G., Mateo, M., Olszewski, E. W., Sen, B., & Woodroofe, M. 2009b, 1475 AJ, 137, 3109 Sales, L. V., Helmi, A., & Battaglia, G. 2010, Adv. Astron., 2010, 194345 Walker, M. G., Olszewski, E. W., & Mateo, M. 2015, MNRAS, 448, 2717 Saviane, I., Held, E. V., & Piotto, G. 1996, A&A, 315, 40 Wetzel, A. R., Deason, A. J., & Garrison-Kimmel, S. 2015, ApJ, 807, 49 Savino, A., Tolstoy, E., Salaris, M., Monelli, M., & de Boer, T. J. L. 2019, A&A, Wetzel, A. R., Hopkins, P. F., Kim, J.-H., et al. 2016, ApJ, 827, L23 630, A116 Wheeler, C., Pace, A. B., Bullock, J. S., et al. 2017, MNRAS, 465, 2420 Sawala, T., Scannapieco, C., Maio, U., & White, S. 2010, MNRAS, 402, Wolf, J., Martinez, G. D., Bullock, J. S., et al. 2010, MNRAS, 406, 1220 1599 Zoccali, M., Gonzalez, O. A., Vasquez, S., et al. 2014, A&A, 562, A66

A152, page 15 of 24 A&A 635, A152 (2020)

Appendix A: Line-of-sight velocity and metallicity A.2. Comparison between measurements of different measurements – consistency checks and datasets obtained with our reduction comparisons We compared the measurements obtained from the P91 dataset A.1. Consistency checks with those from our reduction of the P69 and FLAMES datasets, using the targets they had in common. This was an important We performed consistency checks in order to unveil the pres- step to ensure we could combine the velocity catalogs. ence of possible systematic errors in our velocity determina- The P91 and P69 FORS2 datasets have five targets in tions using fxcor, or other issues. For all the datasets analyzed, common with measured velocities and metallicities, while the we searched for systematics by shifting the template spectrum FLAMES dataset has seven targets in common (four taking those at several velocities (from −50 km s−1 to 500 km s−1 at step of with S/N > 10 Å−1) with the P91 FORS2 dataset and 18 (15) 50 km s−1) and cross-correlating them with the templates them- with the P69 one (see the top row of Fig. A.1 for the velocity selves at rest, through the full wavelength range of data, and also comparison). only around the CaT. We did not find any significant systematic The agreement between these common targets is excellent, shift introduced by the cross-correlation procedure as a function mostly within 1-σ, except for three targets in common between of the velocity shift nor the cross-correlation range. the FLAMES dataset and the P69 FORS2 one (marked as object We also verified that the random errors were well-treated by 1, 2 and 3 in the top row of Fig. A.1) with velocities over 3-σ the fxcor task and if they led to reliable velocity errors. In this from each other. Of these targets, two are in common with the case, we performed a Monte Carlo analysis, randomly adding P91 FORS2 dataset (objects 1 and 3), which are in good agree- the error-spectra from targets having measured S/N of 10, 15, ment with the FLAMES measurements (Fig. A.1 top row, central 20, 25, and 30 Å−1 for the FORS2 datasets, and of 5, 10, 20, 50, panel), but not with those of the P69 FORS2 dataset (Fig. A.1 and 100 Å−1 for the FLAMES one, to the shifted templates of the top row, right panel). We then found that the problem may reside previous step. We performed 250 realizations for each case and with the measurements from the P69 FORS2 catalog. Therefore, found that the procedure tends to produce velocity errors that we inspected these outliers, namely slits 9 and 10 of chip-1 and are underestimated with respect to the velocity scatter from the slit 13 of chip-2 (respectively object 1, 2, and 3 in the cited fig- individual Monte Carlo runs for S/N lower than 10 Å−1, for both ures), finding that slits 9 and 13 have particularly noisy sky-line datasets. Therefore, we take these values as the limit S/N above residuals around the first two CaT lines that may have compro- which we will trust the velocity error estimations. At this S/N, mised their measurements, while the spectrum of slit 10 was the velocity errors were found to be ∼±10 km s−1 in both cases. clean. However, this target was marked as a double star and with We further checked the choice of using the two reddest slit mis-centering problems in F09. We then decided to exclude lines of the CaT for the cross-correlation measurements of the these values from the P69 FORS2 catalog, but to maintain them FLAMES dataset (see previous section), creating two more cat- in the P91 FORS2 and FLAMES datasets. In conclusion, after a alogs of velocity estimations using both the entire CaT range homogeneous analysis, the measurements for stars in common, and the two bluest lines. Cross-correlating between catalogs and which were taken during several epochs as well as with differ- selecting those targets with an S/N > 10 Å−1, which reduced ent instruments and spectral resolution in the case of FORS2 and our sample to 76 objects out of 164, we found very good agree- FLAMES, are found to be in very good agreement. ment, excluding eight targets that had velocity measurement dif- ferences more than 3-σ from each other. We visually checked A.3. Comparison with other works the corresponding spectra for these targets and found that for the outliers the first and/or third line of the CaT resulted affected by Comparison with F09: The l.o.s. velocity and metallicity dis- sky residuals or were hidden by noise, while in general, the cen- tributions of the P91 and F09 datasets show a difference of tral line resulted more clearly visible. For fiber 118, the opposite ∼10 km s−1 and ∼0.25 dex in the mean values, respectively. We happened, and we considered the measurement obtained using investigated whether clear offsets could be found by comparing the full CaT range for this case. the measurements of the targets in common between the two Considering once again the velocities from targets having an studies. There are only three such stars, for which the metallic- S/N >10 Å−1, we looked for cases were the cross-correlation pro- ities agree at the 1-σ level, but the l.o.s. velocities differ at 3-σ cedure did not get a velocity solution, finding only two cases level for two stars, and 1.5-σ for the other one, respectively. The (fibers 18 and 55). Their visual inspection did not show obvious stars discrepant at the 3-σ level were slits 9 and 13 of chip-1 and CaT lines, and we thus decided to discard these two targets. -2, respectively (targets 7 and 18 in F09), whose P91 spectra do Finally, we checked the internal accuracy of our P91 FORS2 not show any particular issue. dataset by inspecting the four targets in common between the Since we could not pinpoint the sources of these differences two pointings. Of these, three had reliable measurements of in the line-of-sight velocities, we decided to re-reduce and rean- velocity and metallicity that were compatible within 1-σ, while alyze the F09 FORS2 dataset, so as to be homogeneous with the the other one did not have reliable measurements. In fact, it did treatment of the P91 data. not show a visible CaT, and its I-magnitude and color were also The comparison of the l.o.s. velocities and metallicities not compatible with the RGB of Tucana (see also Sect.4). We derived from our reduction of the P69 FORS2 dataset and those then decided to discard this target from the sample and to aver- published by F09 showed average offsets of 7 km s−1 and 0.2 dex age the measurements from the other three targets together. in the velocity and metallicity measurements, of the same signs We further excluded two more targets from the P91 FORS2 and comparable to those previously found between the targets dataset since their measurements were not reliable due to high in common between the P91 FORS2 dataset and the F09 cata- residuals in their spectra: these are slit 14, chip-2 of Tuc0 field, log (see Fig. A.1 bottom row, left panel for the velocity com- and slit 17, chip-1 of Tuc1. The high residuals in these two spec- parison). It is unclear what leads to the measured difference in tra were due to the stellar trace falling on a bad CCD-row and a velocity. For the metallicities, the [Fe/H]-CaT EW calibration in poor sky-subtraction that could not be improved, respectively. F09 is equivalent to that used here in the metallicity and mag-

A152, page 16 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph

P91-FORS2 vs P69-FORS2 Flames vs P91-FORS2 Flames vs P69-FORS2 60 60 60

40 40 40 ]

s 20 20 20 /

m 3 k

[ 0 0 1 0

l e h

V 20 20 20 3 2 40 3 1 40 40 1

60 60 60 0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250 F09 vs P69-FORS2 G19 vs P69-FORS2 G19 vs Flames 100 200 300

75 150 200 50 100 ]

s 100 / 25 50 m k

[ 0 0 0

l e h

V 25 50 100 50 100 200 75 150

100 200 300 0 50 100 150 200 250 0 50 100 150 200 250 100 0 100 200 300 Vhel [km/s] Vhel [km/s] Vhel [km/s]

Fig. A.1. Velocity comparison for common targets between datasets – vhel vs. ∆vhel. Top: between the datasets from our internal reduction – left: FORS2 datasets; center: flames vs. P91-FORS2; right: flames vs. P69-FORS2. Red circles indicate the problematic targets highlighted in the main text. Bottom: between our datasets and those from the literature – left: F09 vs. P69-FORS2; central: G19 vs. P69-FORS2; right: G19 vs. flames. nitude regime of Tucana’s stars, and the 0.1 mag difference in S/N > 10 Å−1), in particular at high velocities (>100 km s−1), the adopted VHB value does not have a significant impact on the and there might even be a dependence as a function of velocity final [Fe/H] values. The main source of difference is likely to between the two datasets. We also confirmed that these problems reside in how the EWs were calculated: in F09, Gaussian profiles remain when considering only those targets marked as likely were used to fit the CaT lines, while we adopt Voigt profiles. We members of G19. checked that if we were to adopt Gaussian profiles, we would A systemic velocity for Tucana around 200 km s−1 was −1 obtain lower metallicities of a difference comparable to the sys- found by G19. At vhel,G19 ∼ 200 km s , the comparison in tematic shift we previously found. We want to stress, however, Fig. A.1 (bottom row, right panel) gives velocities around that the Voigt profiles represent a better fit to the CaT lines, in 170−190 km s−1 for our FLAMES measurements of the same particular along the line wings (see e.g., Starkenburg et al. 2010). stars (i.e., velocities close to the systemic velocity we find for Comparison with G19: We compared our own reduced Tucana), as well as values between 0 and 50 km s−1, so, typical datasets with the full catalog of observed targets by G19. There of foreground stars. Our suspicion is that the G19 catalog con- was not a good agreement in all cases. The P91 FORS2 dataset tains an excess of stars with velocities around 200 km s−1. was the one with the lowest number of common targets (6), We investigate this possibility in two ways: by comparing the which showed mainly a systematic offset of ∼30 km s−1. One radial distribution of the G19 members with that obtained from target, however, showed a velocity difference of ∼130 km s−1. photometric observations of Tucana’s stars (test A) and compar- We visually checked for it in our dataset (target 13 of chip-1 in ing the G19 l.o.s. velocity distribution for the stars they con- Tuc0), finding a clean, high S/N spectrum, with no features that sidered as nonmembers with the expectations from a Galactic could have biased the velocity measurement. foreground model (test B). In both cases, we took into account The P69 FORS2 dataset proved to have 16 targets in com- the displacement of the FLAMES pointings from the center of mon with G19 (see Fig. A.1 bottom row, central panel). While Tucana. for a few targets there is good agreement, overall, there is a large Test A: We looked at the normalized cumulative radial distri- scatter in the distribution of velocity differences with one sig- bution of G19’s member stars, and compared it to that obtained nificant outlier (target 1 of chip-2). An inspection of its corre- from the observed surface density profile of Tucana that we mea- sponding spectrum in the P69 data showed the CaT lines shifted sured from the VLT/VIMOS photometric catalog introduced in with respect to the rest frame by ∼200 km s−1, compatible with Sect.4. Although the VLT /VIMOS catalog does not cover the what we found in the cross-correlation measurements, but sig- entire area of the FLAMES pointings, it was sufficient for fol- nificantly different from the ∼25 ± 14 km s−1 reported by G19. lowing the surface-density profile of the RGB stars of Tucana up The comparison between the velocities we derived from our to its nominal tidal radius. treatment of the FLAMES dataset was even more puzzling, con- We found that the number of G19’s member stars in the outer sidering that the values obtained come from the same sample. parts of Tucana, up to where VLT/VIMOS photometry extends, As can be seen from the bottom-right panel of Fig. A.1, there tends to be overestimated with respect to that expected from the is a significant scatter among the 59 targets in common (we photometry for Tucana’s RGB stars in the same area. It could, compared only those measurements from our dataset with a, however, be argued that the FLAMES/GIRAFFE fiber set-up

A152, page 17 of 24 A&A 635, A152 (2020) might imprint a different distribution than that expected from the 2.5 photometry. Therefore, we performed the following “Test B”, Dutton & Macciò (2014) which is instead free from this possible issue. 2.0 Test B: We now concentrate on the l.o.s. velocity distribu- tion of the stars marked as nonmembers in G19, which should 1.5 only contain Galactic contaminants: this distribution shows two c

−1 0 clear peaks in the velocity histogram, one around 0 km s , and, 1 g

−1 o unexpectedly, the other at ∼200 km s (G19, Fig. 6). Making l 1.0 the reasonable assumption that the nonmembers are mainly fore- ground Galactic contaminants, we compared their velocity dis- tribution with that obtained from the Besançon model (Robin 0.5 et al. 2003) generated in the direction of Tucana over an area equivalent to that of a FLAMES-GIRAFFE pointing, and by opt- 0.0 8.5 9.0 9.5 10.0 10.5 11.0 ing for the Besançon model stars to have similar positions on log10 (Mvir/M ) the CMD as on the FLAMES targets. The velocity distribution of the Besançon model stars showed just a single peak around −1 0 km s , with a smooth decline towards negative and positive 30 40 50 60 70 80 −1 velocities, with a tail extending to 300 km s . We checked if the Vmax [km/s] −1 peak at ∼200 km s for G19’s nonmembers could be explained Fig. B.1. Distribution of NFW halos in virial mass – concentration plane from the distribution expected from the Galactic model. We ran- used to construct circular velocity profiles in Fig.5. Halos are color- domly chose, from the synthetic dataset, a number of stars equal coded by Vmax. to that of the G19 nonmembers, and, over 1000 trials, calcu- lated the number of objects extracted from the Besançon model −1 that would have velocities >150 km s : we never got a num- 102 ber of contaminants as high as that of G19’s nonmembers over 9.0 the same velocity range. This indicates that the number of stars 8.5 with velocities ∼200 km s−1 in the G19 catalog of nonmembers 8.0 )

could likely be overestimated. Performing the same exercise for ] s M / 7.5 / * our FLAMES targets that proved to be nonmembers (i.e., with m 1

10 M k ( [

−1

0 P < 0.05) considering the targets with velocities >120 km s c 7.0 1 V g o

(accounting for the observed shift between the datasets), we l found that the observed number of nonmembers is within the 6.5 Local Group (Kirby et al. 2014) 87% (1.5-σ) of the distribution obtained for the contaminants. 6.0 Cetus (Taibi et al. 2018) We speculate that this excess of velocities around Tucana (Taibi et al. 2020) 5.5 −1 100 ∼200 km s in the G19 dataset can be attributed to a sky sub- 102 103 traction problem around the CaT region. In fact, the sky-lines at R1/2 [pc] 8504 Å and 8548 Å, which are around the first and second lines of the CaT, if badly subtracted could lead to strong absorption 20 40 60 80 residuals. These features, in a low-S/N regime and during the c cross-correlation procedure, could be mistaken for the first two 102 CaT lines shifted at ∼200 km s−1, so around the value of the sys- temic velocity reported for Tucana by G19. ] s

Appendix B: Supplementary material from the /

m 101 k [

kinematic analysis c V B.1. Supplementary tables We provide here supplementary tables with results from the Local Group (Kirby et al. 2014) Cetus (Taibi et al. 2018) structural analysis of Tucana using the VLT/VIMOS photomet- Tucana (Taibi et al. 2020) 100 ric dataset, as reported in Sect.4, along with the results from the 102 103 mock tests presented in Sect. 5.1. R1/2 [pc] In all tables, the reported values of the parameters obtained from a Bayesian analysis represent the median of the corre- 9.0 9.5 10.0 10.5 sponding marginalized posterior distributions, with 1-σ errors log10 (Mvir/M ) set as the confidence intervals around the central value enclosing 68% of each distributions. Fig. B.2. Same as Fig. 5, but with halo circular velocity profiles color- coded by concentration (top) and virial mass (bottom). Cetus and Tucana are shown as a triangle and a square, respectively. Top panel: B.2. CDM halo profiles the Local Group dwarf galaxies are color-coded according to their stel- lar mass; a probable correlation is visible between the circular velocity The NFW circular velocity curves shown in Fig.5 are generated and the stellar mass. Bottom panel: the color-coded dots that follow for 300 mock halos, which are uniformly sampled in virial mass the same color schema as the halo virial mass profiles are Local Group between 8.5 ≤ logMvir ≤ 11, and follow the redshift zero halo dwarf galaxies, for which virial masses were obtained from dynamical mass concentration relation from Dutton & Macciò(2014) with modeling (Read & Erkal 2019; Leaman et al. 2012).

A152, page 18 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph

Table B.1. Structural parameters from MCMC analysis fitting an exponential density profile to the RGB population selected in the VLT/VIMOS photometry.

σ0 (α0, δ0) r0 PA  σc (stars arcmin−2) (deg) (arcmin) (deg) (stars arcmin−2) +32 475−30 340.4589 ± 0.0006, −64.4198 ± 0.0004 0.72 ± 0.03 95.6 ± 1.9 0.46 ± 0.02 1.1 ± 0.2 Notes. We report the value of the central density, the coordinates of the optical center, the scale parameters projected along the optical major axis, the position angle of the optical major axis, the ellipticity and the density of constant contamination.

Table B.2. Mock results for the P91 + P69 + FLAMES (P > 0.05) catalog using an input linear rotation model.

k θk Vsys σv k θk Vsys σv ln Brot,disp (km s−1/0)(◦) (km s−1) (km s−1) (km s−1/0)(◦) (km s−1) (km s−1) +1.26 +1.11 +0.88 +9.17 +1.22 +1.33 +5.43 97 0.03−1.22 6.08−1.26 11.40−0.77 96.83−9.18 2.69−1.25 16.56−1.17 34.06−4.78 +1.18 +1.17 +1.61 +6.05 +1.11 +1.30 +5.36 11.2 142 0.00−1.27 6.03−1.18 11.10−1.58 141.36−6.81 0.20−1.27 12.05−1.24 19.67−4.55 − +1.24 +1.14 +2.01 +5.55 − +1.16 +1.14 +4.49 187 0.07−1.29 6.12−1.36 10.84−1.80 187.55−4.70 2.36−1.23 9.70−1.21 9.92−3.97 +1.17 +1.11 +0.84 +12.62 +1.09 +1.19 +4.84 97 0.08−1.28 6.00−1.30 7.73−0.87 96.03−13.37 1.76−1.28 11.63−1.26 19.11−4.55 +1.08 +1.17 +1.75 +9.21 +1.13 +1.18 +4.43 7.5 142 0.00−1.13 5.97−1.26 7.35−1.57 141.11−12.74 0.06−1.03 9.06−1.28 9.26−4.02 − +1.12 +1.21 +1.92 +8.22 − +1.06 +1.16 +3.56 187 0.11−1.18 5.98−1.26 7.01−1.91 187.18−7.34 1.55−1.11 7.85−1.24 3.12−2.90 +1.16 +1.14 +0.91 +16.68 +1.15 +1.12 +4.75 97 0.01−1.21 6.10−1.23 5.94−0.95 96.35−17.49 1.19−1.10 9.59−1.31 10.85−3.87 − +1.30 +1.14 +1.47 +11.33 +1.26 +1.10 +3.54 5.6 142 0.03−1.22 6.01−1.21 5.46−1.54 139.47−16.20 0.04−1.14 7.88−1.20 4.41−3.29 − +1.22 +1.16 +1.97 +12.05 − +1.14 +1.13 +2.92 187 0.17−1.25 6.10−1.25 5.03−2.15 188.44−10.14 1.19−1.18 7.18−1.25 0.41−2.30 +1.17 +1.12 +0.95 +25.01 +1.07 +1.20 +3.42 97 0.04−1.21 6.07−1.28 4.08−0.92 94.80−23.97 0.81−1.13 7.67−1.15 3.85−3.13 − +1.26 +1.16 +1.46 +17.43 +1.17 +1.27 +3.04 3.7 142 0.10−1.18 6.03−1.23 3.50−1.27 139.67−27.83 0.01−1.15 6.88−1.20 0.15−1.99 − +1.13 +1.17 +2.10 +18.70 − +1.06 +1.17 − +1.98 187 0.13−1.23 6.00−1.19 2.87−1.58 188.08−15.21 0.73−1.17 6.55−1.21 1.68−1.24 +1.12 +1.06 +1.11 +45.31 +1.06 +1.18 − +2.48 97 0.07−1.23 6.10−1.31 2.25−1.06 97.21−50.07 0.47−1.21 6.56−1.24 1.40−1.46 − +1.13 +1.14 +1.47 +26.47 − +1.08 +1.17 − +1.53 1.9 142 0.08−1.10 6.00−1.28 1.63−1.23 152.44−46.04 0.04−1.13 6.21−1.14 2.41−0.79 − +1.09 +1.10 +1.74 +31.00 − +1.05 +1.11 − +1.42 187 0.05−1.14 6.01−1.35 1.26−1.13 183.09−27.01 0.32−1.11 6.16−1.17 2.81−0.52 +1.10 +1.19 +1.18 +30.94 +1.07 +1.15 − +0.80 97 0.01−1.16 5.97−1.17 0.17−1.48 146.72−116.54 0.01−1.12 6.03−1.16 3.13−0.28 − +1.09 +1.06 − +1.52 +29.45 − +1.05 +1.12 − +0.89 0.0 142 0.02−1.18 5.97−1.24 0.18−1.24 168.49−48.65 0.03−1.16 6.02−1.23 3.09−0.33 +1.20 +1.16 +1.38 +42.76 +1.14 +1.16 +0.85 187 0.01−1.20 5.91−1.24 −0.04−1.34 176.81−30.81 0.02−1.20 5.98−1.27 −3.11−0.31

Notes. The first two columns represent the input parameters of the rotation model. In all cases, the input systemic velocity Vsys and velocity −1 dispersion σv were fixed to 0 and 6 km s , respectively. The middle columns are the recovered parameter fitting a linear rotation model and a dispersion-only one, respectively. The last column is the Bayes factor accounting for the evidences of the two models.

scatter of σlnc = 0.25. These halos are shown in Fig. B.1, color- is visible between the circular velocity and the stellar mass. coded by their Vmax. In the bottom panel, observed dwarf galaxies with virial In Fig. B.2, to be comprehensive, we reproduce plots where masses estimated from dynamical modeling (Read & Erkal the mock halos are color-coded by concentration, or virial mass. 2019; Leaman et al. 2012) are also color-coded, illustrat- In the top panel, the Local Group dwarf galaxies are color- ing the tension between the predicted and observed density coded according to their stellar mass; a probable correlation profiles.

A152, page 19 of 24 A&A 635, A152 (2020)

Appendix C: Additional tables

Table C.1. Properties of observed P91-FORS2 dataset in the Tucana dSph.

Field Slit RA, Dec (J2000) V ± δVI ± δI vhel ± δvhel [Fe/H] ± δ[Fe/H] S/NP −1 −1 [deg] [km s ] [dex] [pxl ] P1 P2 P3 0 1 340.4913, −64.41896 22.79 ± 0.02 21.44 ± 0.04 176.6 ± 7.1 −1.37 ± 0.16 22.5 0.99 0.99 0.99 0 2 340.4834, −64.41816 22.75 ± 0.03 21.53 ± 0.04 157.8 ± 6.6 −2.00 ± 0.11 21.4 0.99 0.99 0.99 0 3 340.4762, −64.41057 22.33 ± 0.02 21.19 ± 0.04 189.8 ± 8.3 −2.66 ± 0.14 29.4 0.99 0.99 0.99 0 4 340.4645, −64.42837 22.69 ± 0.02 21.33 ± 0.04 183.9 ± 5.5 −2.17 ± 0.11 24.5 0.99 0.99 0.99 0 5 340.4565, −64.42895 22.88 ± 0.02 21.56 ± 0.03 167.0 ± 10.8 −1.75 ± 0.14 19.4 0.99 0.99 0.99 0 6 340.4511, −64.42363 22.19 ± 0.02 20.55 ± 0.03 165.5 ± 11.7 −1.25 ± 0.07 12.2 0.99 0.99 0.99 0 7 340.4421, −64.42360 22.29 ± 0.02 20.87 ± 0.03 183.6 ± 4.1 −1.29 ± 0.09 38.3 0.99 0.99 0.99 0 8 340.4334, −64.42592 22.81 ± 0.03 21.57 ± 0.03 182.8 ± 4.9 −1.37 ± 0.17 16.2 0.99 0.99 0.99 0 9 340.4283, −64.41176 22.72 ± 0.02 21.46 ± 0.04 155.1 ± 10.0 −2.12 ± 0.18 21.3 0.99 0.99 0.99 0 10 340.4199, −64.41700 22.83 ± 0.02 21.49 ± 0.04 174.2 ± 5.5 −1.28 ± 0.16 23.1 0.99 0.99 0.99 0 11 340.4090, −64.41599 22.34 ± 0.02 20.91 ± 0.03 183.6 ± 5.7 −1.87 ± 0.10 36.3 0.99 0.99 0.99 0 12 340.3984, −64.41116 22.44 ± 0.02 20.96 ± 0.04 168.7 ± 4.4 −1.52 ± 0.11 29.0 0.99 0.99 0.99 0 13 340.3879, −64.41004 22.89 ± 0.02 21.55 ± 0.04 167.6 ± 5.5 −1.29 ± 0.18 20.7 0.99 0.99 0.99 0 14 340.3760, −64.42418 22.67 ± 0.02 21.36 ± 0.04 187.4 ± 5.0 −1.84 ± 0.14 26.6 0.99 0.99 0.99 0 15 340.3693, −64.41906 22.89 ± 0.02 21.82 ± 0.04 178.5 ± 10.2 −1.92 ± 0.18 18.8 0.99 0.99 0.99 0 16 340.3576, −64.41379 22.39 ± 0.02 20.17 ± 0.03 43.3 ± 5.8 −2.05 ± 0.05 54.0 Ph Rep. 340.6158, −64.40893 22.96 ± 0.02 21.36 ± 0.04 146.1 ± 7.0 −1.88 ± 0.12 27.0 0.02 0.02 0.03 0 31 147.5 ± 13.6 −1.81 ± 0.18 25.2 1 31 145.6 ± 8.3 −1.94 ± 0.17 28.9 Rep. 340.6056, −64.4071 21.65 ± 0.02 21.29 ± 0.03 // 24.3 Ph 0 32 // 22.7 1 32 // 25.9 Rep. 340.5832, −64.43912 22.95 ± 0.02 21.36 ± 0.03 225.8 ± 4.4 −1.67 ± 0.09 26.5 0.00 0.00 0.00 0 33 221.8 ± 7.2 −1.75 ± 0.12 24.3 1 33 228.3 ± 5.6 −1.55 ± 0.15 28.6 0 34 340.5744, −64.41475 23.63 ± 0.03 22.46 ± 0.05 // 7.7 Ph Rep. 340.5692, −64.38342 22.39 ± 0.02 20.62 ± 0.03 179.8 ± 4.7 −1.83 ± 0.07 35.0 0.89 0.88 0.88 0 35 181.0 ± 6.4 −1.88 ± 0.11 32.5 1 36 178.3 ± 6.9 −1.79 ± 0.09 37.5 0 36 340.5508, −64.43098 22.78 ± 0.02 21.50 ± 0.04 185.0 ± 6.2 −1.49 ± 0.16 20.2 0.99 0.99 0.99 0 37 340.5418, −64.41572 23.23 ± 0.02 22.03 ± 0.05 175.0 ± 6.8 −1.62 ± 0.20 14.7 0.99 0.99 0.99 0 38 340.5302, −64.43026 23.65 ± 0.03 22.63 ± 0.08 182.8 ± 15.3 −1.40 ± 0.50 8.2 0.99 0.99 0.99 0 39 340.5278, −64.40337 23.15 ± 0.02 21.92 ± 0.04 185.3 ± 13.4 −1.32 ± 0.27 16.1 0.99 0.99 0.99 0 40 340.5186, −64.41192 22.90 ± 0.03 21.58 ± 0.05 162.1 ± 7.3 −1.08 ± 0.25 21.6 0.99 0.99 0.99 0 41 340.5072, −64.42974 22.39 ± 0.02 20.84 ± 0.03 179.2 ± 5.0 −1.18 ± 0.10 30.8 0.99 0.99 0.99 1 1 340.5224, −64.43385 23.46 ± 0.03 22.25 ± 0.07 214.3 ± 18.2 −1.79 ± 0.31 9.6 0.98 0.99 0.99 1 2 340.5116, −64.44016 22.67 ± 0.02 21.36 ± 0.05 170.2 ± 5.1 −2.02 ± 0.13 23.7 0.99 0.99 0.99 1 3 340.5134, −64.41830 22.78 ± 0.02 21.44 ± 0.05 177.5 ± 2.9 −1.25 ± 0.16 26.6 0.99 0.99 0.99 1 4 340.4990, −64.42399 22.63 ± 0.02 21.33 ± 0.03 179.4 ± 4.2 −2.15 ± 0.17 26.6 0.99 0.99 0.99 1 5 340.4922, −64.42342 22.63 ± 0.02 21.41 ± 0.04 192.5 ± 6.9 −1.78 ± 0.10 26.6 0.99 0.99 0.99 1 6 340.4890, −64.41446 23.31 ± 0.02 22.12 ± 0.06 221.8 ± 13.8 −1.81 ± 0.35 13.2 0.98 0.98 0.98 1 7 340.4818, −64.41523 22.75 ± 0.02 21.33 ± 0.04 176.7 ± 4.0 −0.94 ± 0.16 27.9 0.99 0.99 0.99 1 8 340.4706, −64.42798 22.86 ± 0.02 21.54 ± 0.05 185.4 ± 6.8 −0.89 ± 0.20 22.0 0.99 0.99 0.99 1 9 340.4658, −64.42591 22.78 ± 0.02 21.49 ± 0.03 174.6 ± 4.5 −1.06 ± 0.19 20.3 0.99 0.99 0.99 1 10 340.4619, −64.41730 22.60 ± 0.02 21.29 ± 0.03 185.8 ± 3.7 −1.01 ± 0.19 25.6 0.99 0.99 0.99 1 11 340.4553, −64.41642 22.42 ± 0.02 21.12 ± 0.03 188.1 ± 5.0 −1.48 ± 0.14 33.6 0.99 0.99 0.99

Notes. Column (1) field Tuc0 and Tuc1; (2) slit aperture: numbers <30 indicate observed targets from chip-1, otherwise from chip-2 – numbers counted from bottom to top of the CCD; (3) RA-Dec coordinated in J2000; (4) V band magnitude with error from VLT/FORS2 photometric catalog; (5) I band magnitude with error from same catalog; (6) l.o.s. heliocentric velocity with error; (7) metallicity with error; (8) S/N in pxl−1 (the conversion factor to Å−1 is 1.09); (9) probability of membership – the three columns indicate the probabilities obtained using the P91 dataset alone (P1), combining with the P69 (P2) and further adding the FLAMES data (P3). Four stars had repeated measurements: targets 31, 32, 33, and 35 from Tuc0 field corresponding to targets 31, 32, 33, and 36 from Tuc1 field, respectively; in this table, we report the single measurements as well as the averaged values used during the analysis process. In the P columns, stars marked as “Ph” are nonmembers, excluded according to their magnitudes and colors, or because they lack reliable measurements (see Sect.4 for full description).

A152, page 20 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph

Table C.1. continued.

Field Slit RA, Dec (J2000) V ± δVI ± δI vhel ± δvhel [Fe/H] ± δ[Fe/H] S/NP −1 −1 [deg] [km s ] [dex] [pxl ] P1 P2 P3 1 12 340.4446, −64.42166 22.73 ± 0.03 21.47 ± 0.03 189.7 ± 5.5 −1.30 ± 0.18 22.3 0.99 0.99 0.99 1 13 340.4351, −64.42136 22.59 ± 0.02 21.24 ± 0.03 185.1 ± 9.4 −1.47 ± 0.16 26.2 0.99 0.99 0.99 1 14 340.4321, −64.41439 22.61 ± 0.02 21.15 ± 0.03 177.0 ± 5.5 −1.18 ± 0.17 27.7 0.99 0.99 0.99 1 15 340.4200, −64.42276 22.53 ± 0.02 21.24 ± 0.04 189.5 ± 8.9 −2.89 ± 0.17 24.7 0.99 0.99 0.99 1 16 340.4267, −64.39091 22.43 ± 0.02 20.59 ± 0.03 −8.5 ± 4.1 −1.88 ± 0.10 48.0 0.00 0.00 0.00 1 17 340.4098, −64.41052 23.05 ± 0.02 21.79 ± 0.05 // 16.1 Ph 1 18 340.4026, −64.40694 22.67 ± 0.02 21.35 ± 0.04 172.9 ± 5.1 −1.89 ± 0.17 25.9 0.99 0.99 0.99 1 19 340.3919, −64.41256 22.70 ± 0.02 21.36 ± 0.04 175.5 ± 5.4 −1.95 ± 0.13 23.7 0.99 0.99 0.99 1 20 340.3787, −64.43124 22.22 ± 0.02 22.19 ± 0.05 // 7.9 Ph 1 34 340.5865, −64.38883 20.16 ± 0.01 19.47 ± 0.02 // 49.0 Ph 1 35 340.5724, −64.39045 23.15 ± 0.03 20.41 ± 0.04 // 27.3 Ph 1 37 340.5571, −64.38788 22.93 ± 0.02 21.28 ± 0.05 154.9 ± 4.5 −2.08 ± 0.16 29.1 0.06 0.07 0.08

Table C.2. Properties of observed P69-FORS2 dataset in the Tucana dSph.

Slit RA, Dec (J2000) V ± δVI ± δI vhel ± δvhel [Fe/H] ± δ[Fe/H] S/NP −1 −1 [deg] [km s ] [dex] [pxl ] P1 P2 P3 1 340.4472, −64.42209 22.44 ± 0.02 20.97 ± 0.02 180.4 ± 4.8 −1.07 ± 0.17 20.4 0.99 0.99 0.99 2 340.4532, −64.42234 22.48 ± 0.02 21.26 ± 0.02 176.3 ± 6.8 −1.56 ± 0.29 15.0 0.99 0.99 0.99 3 340.4586, −64.39465 22.22 ± 0.02 20.96 ± 0.02 144.2 ± 5.2 −1.52 ± 0.22 19.8 0.17 0.01 0.02 4 340.4647, −64.4236 22.64 ± 0.01 21.22 ± 0.01 186.8 ± 7.4 −1.61 ± 0.16 19.4 0.99 0.99 0.99 5 (a) 340.4698, −64.43846 23.26 ± 0.03 21.50 ± 0.05 130.4 ± 10.1 −1.37 ± 0.20 11.9 0.13 0.04 0.04 6 340.4748, −64.41809 22.39 ± 0.01 20.87 ± 0.01 184.8 ± 4.6 −1.96 ± 0.10 29.8 0.99 0.99 0.99 7 340.4799, −64.40572 22.28 ± 0.02 20.85 ± 0.02 178.9 ± 4.8 −1.92 ± 0.17 24.5 0.99 0.99 0.99 8 340.4881, −64.41732 22.73 ± 0.03 21.19 ± 0.03 177.7 ± 6.2 −1.64 ± 0.13 18.9 0.99 0.99 0.99 9 (b) 340.4920, −64.42341 22.61 ± 0.02 21.35 ± 0.02 232.8 ± 8.0 −1.28 ± 0.22 15.4 Ph 10 340.4963, −64.41156 22.45 ± 0.02 21.01 ± 0.02 225.0 ± 7.2 −2.00 ± 0.18 19.8 Ph 11 340.5012, −64.41237 22.73 ± 0.01 21.49 ± 0.01 166.8 ± 5.4 −1.19 ± 0.35 13.0 0.99 0.99 0.99 12 340.5069, −64.40718 22.28 ± 0.02 20.86 ± 0.02 181.8 ± 8.3 −1.87 ± 0.18 21.8 0.99 0.99 0.99 13 340.5105, −64.42116 22.31 ± 0.02 20.83 ± 0.02 209.9 ± 7.6 −1.96 ± 0.16 18.3 0.99 0.96 0.96 14 340.5140, −64.41146 22.49 ± 0.03 20.96 ± 0.02 191.7 ± 5.3 −1.41 ± 0.14 23.9 0.99 0.99 0.99 23 (b) 340.5695, −64.3833 22.41 ± 0.03 20.65 ± 0.04 179.5 ± 5.7 −1.95 ± 0.16 25.2 0.79 24 340.5742, −64.37577 0.0 ± 0.0 0.0 ± 0.0 4.2 ± 4.8 / 32.8 0.00 0.00 0.00 25 (a),(b) 340.5836, −64.43902 22.95 ± 0.02 21.46 ± 0.03 236.4 ± 14.1 −2.01 ± 0.26 17.4 0.63 31 340.3294, −64.4244 21.78 ± 0.02 20.59 ± 0.02 208.6 ± 5.2 / 23.5 Ph 33 (a) 340.3435, −64.39041 22.61 ± 0.02 21.33 ± 0.02 110.3 ± 14.9 −2.42 ± 0.28 14.6 0.00 0.00 0.00 35 340.3536, −64.41425 22.59 ± 0.02 21.0 ± 0.02 76.6 ± 7.3 −1.77 ± 0.15 21.4 0.00 0.00 0.00 39 (a),(b) 340.3762, −64.4242 22.69 ± 0.02 21.38 ± 0.02 197.3 ± 8.3 −2.11 ± 0.21 12.9 0.99 40 340.3820, −64.45349 0.0 ± 0.0 0.0 ± 0.0 25.6 ± 11.1 / 27.0 0.00 0.00 0.00 43 (b) 340.3973, −64.41121 22.45 ± 0.02 20.98 ± 0.02 207.8 ± 7.3 −1.14 ± 0.17 21.3 Ph 44 340.4024, −64.41617 22.27 ± 0.02 20.77 ± 0.02 180.1 ± 5.2 −2.20 ± 0.12 25.7 0.99 0.99 0.99 45 340.4100, −64.40206 22.41 ± 0.02 21.03 ± 0.02 174.9 ± 5.9 −2.12 ± 0.15 21.3 0.99 0.99 0.99 46 (a) 340.4163, −64.40805 21.09 ± 0.01 20.36 ± 0.01 −160.6 ± 4.8 / 31.7 Ph 47 340.4206, −64.41588 22.68 ± 0.02 21.28 ± 0.02 195.7 ± 6.0 −1.33 ± 0.20 14.7 0.99 0.99 0.99 49 340.4348, −64.40812 22.37 ± 0.03 21.01 ± 0.02 172.5 ± 6.8 −1.97 ± 0.18 24.3 0.99 0.99 0.99

Notes. Column (1) slit aperture: numbers <30 indicate observed targets from chip-1, otherwise from chip-2 – numbers counted from bottom to top of the CCD; (2) RA-Dec coordinated in J2000; (3) V band magnitude with error from VLT/FORS2 photometric catalog; (4) I band magnitude with error from same catalog; (5) l.o.s. heliocentric velocity with error; (6) metallicity with error; (7) S/N in pxl−1 (the conversion factor to Å−1 is 1.08); (8) probability of membership according to the best-fitting kinematic model – the three columns indicate the probabilities obtained using the P69 dataset alone (P1), combining with the P91 (P2) and further adding the FLAMES data (P3). In the P columns, stars marked as “Ph” are nonmembers, excluded according to their magnitudes and colors, or because their measurements proved to be unreliable (see Sect.4 and AppendixA for full description). (a)New with respect to Fraternali et al.(2009). (b)In common with the P91-FORS2 dataset.

A152, page 21 of 24 A&A 635, A152 (2020)

Table C.3. Properties of observed FLAMES/GIRAFFE dataset in the Tucana dSph.

Object Fiber RA, Dec (J2000) V I vhel ± δvhel S/NP −1 −1 [deg] [km s ] [pxl ] P1 P2 44068 1 340.34104, −64.366778 22.976 20.486 19.7 ± 6.4 12.0 Ph 58039 2 340.27637, −64.447250 23.368 22.183 / 0.8 Ph 53427 3 340.38987, −64.422028 22.861 21.823 / 2.2 Ph 49196 4 340.40287, −64.401111 23.340 22.133 / 3.4 Ph 100020 5 340.32892, −64.424889 24.880 22.406 / 3.3 Ph 51120 6 340.36892, −64.411639 22.996 21.890 160.6 ± 18.8 3.2 Ph 100019 7 340.38075, −64.453639 20.716 19.920 23.6 ± 9.6 15.3 Ph 39027 8 340.54671, −64.333556 22.992 21.378 79.6 ± 18.1 6.2 0.00 0.00 31104 9 340.34300, −64.279222 22.230 20.723 292.6 ± 16.6 9.4 0.00 0.00 27260 10 340.47533, −64.253083 22.898 21.565 39.3 ± 9.0 3.8 Ph 37852 11 340.49483, −64.326306 23.199 21.657 7.0 ± 7.1 5.2 0.0 0.00 37244 12 340.42596, −64.322194 22.103 20.879 191.3 ± 14.3 7.8 0.0 0.00 45927 13 340.13696, −64.379889 22.964 21.509 451.9 ± 4.5 3.7 Ph 44431 14 340.31504, −64.368972 22.904 21.351 −201.6 ± 13.2 4.9 Ph 45504 15 340.21796, −64.376917 23.023 21.510 234.5 ± 13.8 5.4 0.00 0.00 40746 16 340.45733, −64.345333 22.190 21.000 230.3 ± 18.6 6.6 0.00 0.00 30380 17 340.29579, −64.274167 23.089 21.909 / 3.5 Ph 34523 18 340.51400, −64.302361 22.602 21.179 / 5.3 Ph 32778 19 340.45058, −64.290778 24.225 21.797 / 3.8 Ph 32116 20 340.50783, −64.286083 22.204 20.595 32.4 ± 5.2 11.8 0.00 0.00 43472 21 340.44033, −64.363083 22.615 21.174 −38.7 ± 11.9 4.4 Ph 39444 22 340.25721, −64.336389 22.121 20.730 126.7 ± 7.2 10.0 0.00 0.00 24956 23 340.37875, −64.235778 0.000 0.000 / 3.2 Ph 33846 24 340.36354, −64.297972 22.937 21.543 −42.5 ± 10.3 5.3 0.00 0.00 42547 25 340.52042, −64.357194 22.577 21.207 105.9 ± 8.6 4.9 Ph 30098 26 340.50979, −64.272278 17.807 16.986 123.9 ± 1.1 109.6 Ph 20262 27 340.21804, −64.200806 22.595 20.826 36.0 ± 18.1 8.6 0.00 0.00 28726 28 340.55079, −64.263167 23.994 21.447 16.6 ± 18.2 5.7 Ph 28435 29 340.56392, −64.261222 0.000 0.000 9.2 ± 13.2 3.8 Ph 36322 30 340.44375, −64.315444 22.305 20.703 30.3 ± 11.8 10.2 0.00 0.00 40890 31 340.42958, −64.346389 23.053 21.507 −62.0 ± 23.8 5.7 0.00 0.00 22014 32 340.56696, −64.213194 22.229 20.861 −22.4 ± 10.0 7.9 0.00 0.00 24597 33 340.53050, −64.233250 23.047 21.985 / 2.4 Ph 20184 34 340.44471, −64.200333 22.705 20.813 119.3 ± 12.0 6.8 0.00 0.00 25301 35 340.69692, −64.238194 23.488 22.373 / 3.1 Ph 21919 36 340.63321, −64.212528 22.609 21.043 61.9 ± 7.7 6.1 0.00 0.00 31878 37 340.64842, −64.284611 22.696 21.652 / 4.3 Ph 24201 38 340.67017, −64.230417 22.557 21.386 / 2.9 Ph 4032 39 340.50846, −64.087278 22.940 21.389 / 1.5 Ph 16892 40 340.56733, −64.178917 22.773 21.559 / 1.9 Ph 31326 41 340.62975, −64.280972 22.065 20.503 12.7 ± 7.2 13.4 0.00 0.00 5642 42 340.45183, −64.099333 0.000 0.000 169.0 ± 14.0 2.2 Ph 36740 43 340.69904, −64.318389 22.802 21.257 226.6 ± 13.3 5.1 0.00 0.00 13690 44 340.76192, −64.156333 22.987 21.887 / 2.1 Ph 32638 45 340.56775, −64.289861 22.388 20.678 111.2 ± 14.0 12.0 0.00 0.00 32163 46 340.58575, −64.286389 22.559 20.710 117.7 ± 10.7 10.4 0.00 0.00 30991 47 340.65567, −64.278278 18.715 17.675 −50.1 ± 1.0 42.5 Ph 20980 48 340.70929, −64.205528 22.358 20.702 1.2 ± 6.3 10.9 0.00 0.00 18416 49 340.65187, −64.188583 22.400 21.031 30.2 ± 6.2 5.5 0.00 0.00 24848 50 340.61529, −64.235028 22.317 20.877 187.0 ± 40.5 7.5 Ph

Notes. Column (1) object-ID from fits header; (2) fiber-ID used through text; (3) RA-Dec coordinated in J2000; (4) V band magnitude transformed from DES photometric catalog; (5) I band magnitude transformed from the same catalog; (6) l.o.s. heliocentric velocity with error; (7) S/N in pxl−1 (the conversion factor to Å−1 is 2.24); (8) probability of membership according to the best-fitting kinematic model – the two columns indicate the probabilities obtained using the FLAMES dataset alone (P1), and combining with the FORS2 data (P2). In the P columns, stars marked as “Ph” are non-members excluded according to their magnitudes and colors or because their measurements resulted to be not reliable or because having a S/N < 10 Å−1 (see Sect.4 and AppendixA for full description). (a)In common with the FORS2 dataset.

A152, page 22 of 24 S. Taibi et al.: Chemo-kinematics of the Tucana dSph

Table C.3. continued.

Object Fiber RA, Dec (J2000) V I vhel ± δvhel S/NP −1 −1 [deg] [km s ] [pxl ] P1 P2 21528 51 340.70312, −64.209722 22.354 20.615 45.3 ± 3.8 7.0 0.00 0.00 23752 52 340.47312, −64.226722 23.407 20.383 −23.1 ± 8.4 10.2 Ph 31071 53 340.76042, −64.279000 22.194 20.633 −49.6 ± 8.6 10.4 0.00 0.00 31001 54 340.52425, −64.278361 19.060 18.133 −8.5 ± 1.3 48.2 Ph 16300 55 340.81579, −64.175000 22.232 20.746 / 6.2 Ph 19109 56 340.75412, −64.193778 23.158 21.607 10.0 ± 28.4 3.2 Ph 18966 57 340.82242, −64.192750 22.928 21.727 −45.0 ± 11.4 4.0 Ph 32477 58 340.72533, −64.288583 22.329 20.713 166.8 ± 9.7 8.8 0.00 0.00 27434 59 340.42087, −64.254250 22.831 22.089 / 4.0 Ph 23221 60 340.79529, −64.222528 22.896 21.570 / 3.7 Ph 22156 61 340.92625, −64.214333 22.016 20.832 36.2 ± 5.0 4.6 Ph 37211 62 340.60975, −64.322028 23.534 22.110 71.6 ± 8.9 3.9 Ph 35540 63 340.60425, −64.309583 23.512 21.524 35.1 ± 8.0 6.0 0.00 0.00 20829 64 340.86492, −64.204417 22.850 21.214 64.4 ± 9.6 4.8 Ph 35289 65 340.95317, −64.307861 22.275 21.177 150.2 ± 22.5 3.9 Ph 35409 66 340.89879, −64.308722 22.419 20.780 −53.4 ± 5.6 7.1 0.00 0.00 32709 67 340.92150, −64.290306 22.070 20.786 145.1 ± 5.7 8.4 0.00 0.00 28348 68 340.93237, −64.260694 21.678 21.149 74.5 ± 15.9 1.2 Ph 21201 69 340.56067, −64.207028 22.672 20.824 21.8 ± 7.2 6.3 0.00 0.00 34708 70 340.81462, −64.303722 22.380 21.111 31.0 ± 10.6 5.1 0.00 0.00 26258 71 340.91354, −64.245583 22.960 21.634 / 4.3 Ph 40705 72 340.62008, −64.345167 23.701 22.204 / 4.1 Ph 25087 73 340.92037, −64.236667 23.166 21.664 / 3.2 Ph 43260 74 340.79483, −64.361667 21.816 20.598 119.6 ± 6.4 7.1 0.00 0.00 38186 75 340.82146, −64.328167 22.477 21.367 −3.6 ± 10.6 3.1 Ph 44653 76 340.78854, −64.370306 23.135 22.253 247.6 ± 15.6 4.5 Ph 27730 77 340.57954, −64.256472 19.371 17.200 33.2 ± 2.5 25.3 Ph 32288 78 340.60825, −64.287361 21.536 19.870 7.0 ± 3.6 15.0 Ph 43618 79 340.65275, −64.363917 22.909 21.618 18.2 ± 15.2 5.2 0.00 0.00 51984 80 340.82829, −64.415750 22.555 21.163 68.8 ± 9.9 3.5 Ph 47042 81 340.80346, −64.387083 23.206 22.295 32.4 ± 10.2 2.5 Ph 48384 82 340.67171, −64.396111 22.925 21.382 151.5 ± 14.7 5.6 0.11 0.09 40996 83 340.55800, −64.347111 21.789 19.157 8.5 ± 4.5 17.1 Ph 48246 84 340.90804, −64.395111 23.503 22.183 75.9 ± 10.6 1.6 Ph 7717 85 340.36279, −64.114750 23.246 22.057 −127.7 ± 16.4 2.0 Ph 53621 86 340.76875, −64.422778 22.723 21.204 138.3 ± 7.6 4.5 Ph 46156 87 340.69479, −64.381472 23.079 21.690 −78.4 ± 13.8 4.2 Ph 100022 88 340.57396, −64.375972 21.449 20.089 −1.9 ± 3.9 15.0 0.00 0.00 48450 89 340.69687, −64.396500 23.093 22.168 51.1 ± 11.3 1.8 Ph 54894 90 340.66104, −64.428417 23.263 21.835 165.0 ± 11.5 3.5 Ph 100003 91 340.47454, −64.418306 0.000 0.000 178.5 ± 3.6 8.5 0.99 (a) 49630 92 340.52779, −64.403528 23.248 22.008 173.9 ± 9.5 3.2 Ph 100014 93 340.45825, −64.394861 22.315 20.943 145.3 ± 5.4 8.7 0.17 (a) 100009 94 340.49612, −64.411722 0.000 0.000 188.7 ± 3.6 8.6 0.99 0.99 100013 95 340.50742, −64.407194 0.000 0.000 73.3 ± 11.1 1.7 Ph 100021 96 340.56921, −64.383500 22.354 20.657 186.2 ± 12.6 6.8 0.80 (a) 52804 97 340.63075, −64.419389 22.454 21.103 −194.8 ± 5.2 6.5 0.00 0.00 52613 98 340.61704, −64.418556 23.114 21.918 188.8 ± 15.0 3.6 Ph 39325 99 340.55700, −64.335500 23.549 20.742 6.0 ± 12.0 8.8 Ph 100002 100 340.46446, −64.423806 21.317 20.260 188.0 ± 15.6 5.7 0.99 (a) 49958 101 340.42304, −64.405278 22.881 21.748 167.9 ± 12.7 3.4 Ph 100016 102 340.40967, −64.402278 22.405 21.056 174.0 ± 6.1 6.0 0.99 (a) 100005 103 340.43446, −64.408333 22.420 21.021 180.9 ± 5.0 8.0 0.99 (a) 50951 104 340.44454, −64.410722 22.730 21.691 / 2.8 Ph 9224 105 340.31279, −64.126778 22.600 21.255 −122.7 ± 10.2 3.7 Ph 100018 106 340.35333, −64.414500 22.687 20.941 67.1 ± 3.9 7.1 0.00 (a) 100017 107 340.39833, −64.411361 22.249 20.909 173.4 ± 6.0 7.7 0.99 (a)

A152, page 23 of 24 A&A 635, A152 (2020)

Table C.3. continued.

Object Fiber RA, Dec (J2000) V I vhel ± δvhel S/NP −1 −1 [deg] [km s ] [pxl ] P1 P2 100006 108 340.49225, −64.423611 22.349 21.126 191.5 ± 8.2 5.6 0.99 (a) 100001 109 340.44696, −64.422306 21.427 20.323 175.1 ± 3.8 8.4 0.99 (a) 47569 110 340.34308, −64.390611 22.639 21.297 149.9 ± 11.1 5.3 0.63 0.54 46004 111 340.42396, −64.380417 22.728 21.480 18.6 ± 17.8 5.0 0.00 0.00 54972 112 340.25037, −64.428806 23.127 22.017 / 2.2 Ph 43720 113 340.41837, −64.364556 21.006 19.251 63.2 ± 2.5 30.8 Ph 21138 114 340.59517, −64.206583 23.023 21.318 105.6 ± 26.0 6.0 0.00 0.00 39993 115 340.32687, −64.340028 21.844 20.614 −43.2 ± 10.5 6.5 0.00 0.00 50864 116 340.38775, −64.410250 22.784 21.569 / 3.0 Ph 39557 117 340.14258, −64.337139 22.481 21.281 / 2.7 Ph 25558 118 340.42058, −64.240028 22.349 20.730 172.8 ± 8.6 5.9 0.00 0.00 33805 119 340.28633, −64.297639 22.541 21.346 −364.1 ± 7.5 3.3 Ph 34818 120 340.29067, −64.304556 22.643 21.338 / 3.0 Ph 28959 121 340.13192, −64.264889 22.808 21.283 / 2.5 Ph 26205 122 340.18454, −64.245194 22.136 20.540 −1.2 ± 6.2 5.3 0.00 0.00 32371 123 340.25575, −64.287861 22.674 20.963 161.0 ± 13.4 4.5 Ph 26562 124 340.47250, −64.248028 22.589 21.159 −404.5 ± 8.8 4.3 Ph 33199 125 340.05833, −64.293333 23.122 22.000 / 2.7 Ph 38665 126 340.39025, −64.331333 20.370 19.311 100.5 ± 5.4 16.8 Ph 15314 127 340.19367, −64.168111 22.785 21.271 193.8 ± 23.0 2.6 Ph 25202 128 340.42783, −64.237500 23.155 21.857 319.1 ± 12.4 2.7 Ph 22526 129 340.27612, −64.217083 22.241 19.161 6.8 ± 7.8 11.0 Ph 28804 130 340.29137, −64.263722 22.447 21.320 −419.7 ± 3.3 3.0 Ph 22297 131 340.13142, −64.215389 21.895 20.688 −37.1 ± 10.1 5.1 0.00 0.00 25382 132 340.18987, −64.238778 0.000 0.000 / 3.0 Ph 13504 133 340.24646, −64.155139 22.938 21.329 136.9 ± 15.8 3.6 Ph 19228 134 340.57158, −64.194472 23.355 22.151 / 2.9 Ph 5411 135 340.47821, −64.097583 22.284 21.120 / 3.0 Ph 7674 136 340.45483, −64.114306 22.383 21.151 / 3.2 Ph 24957 137 340.75162, −64.235778 22.854 21.301 / 4.2 Ph 24991 138 340.76054, −64.235972 23.274 22.108 / 2.6 Ph 31905 139 340.54608, −64.284778 22.730 21.167 45.5 ± 16.0 5.6 0.00 0.00 20561 140 340.84067, −64.202861 21.969 20.723 199.4 ± 10.7 6.1 0.00 0.00 13652 141 340.89246, −64.156111 22.599 20.829 / 3.2 Ph 29029 142 340.98737, −64.265417 23.359 22.117 / 1.9 Ph 28008 143 340.92996, −64.258472 0.000 0.000 83.4 ± 14.9 4.8 Ph 43939 144 340.83217, −64.366000 22.510 20.554 40.1 ± 5.4 7.1 0.00 0.00 41813 145 340.53267, −64.352583 22.750 21.276 142.6 ± 18.3 6.3 0.01 0.01 40477 146 340.93167, −64.343667 19.724 18.922 168.6 ± 3.4 19.6 Ph 38363 147 340.56947, −64.329389 23.588 21.699 −44.2 ± 9.7 3.2 Ph 54686 148 340.83783, −64.427528 22.108 20.905 180.1 ± 11.6 3.4 Ph 41818 149 340.48062, −64.352611 0.000 0.000 160.9 ± 21.5 3.5 Ph 52574 150 340.77154, −64.418333 23.343 22.174 68.4 ± 5.3 2.0 Ph 50194 151 340.45900, −64.406667 23.010 21.531 / 3.8 Ph 52007 152 340.54187, −64.415889 23.029 21.919 / 2.8 Ph 100015 153 340.51367, −64.411667 22.436 20.912 181.9 ± 6.3 5.2 0.99 (a) 52859 154 340.60392, −64.419611 22.947 21.748 / 2.3 Ph 54691 155 340.63350, −64.427556 23.203 22.063 / 2.5 Ph 60765 156 340.69737, −64.465083 22.639 21.274 147.8 ± 17.3 3.0 Ph 50636 157 340.61592, −64.409111 23.031 21.272 / 3.2 Ph 50350 158 340.45050, −64.407500 22.775 21.648 197.8 ± 9.3 2.6 Ph 46819 159 340.41362, −64.385528 22.469 20.773 −34.9 ± 11.4 6.6 0.00 0.00 100007 160 340.42029, −64.416056 22.412 21.129 −88.3 ± 6.2 4.7 Ph 53700 161 340.46612, −64.423111 0.000 0.000 185.6 ± 11.5 2.5 Ph 51422 162 340.36446, −64.413167 23.310 22.180 204.1 ± 4.6 2.8 Ph 51322 163 340.18804, −64.412694 22.476 21.120 / 3.0 Ph 43254 164 340.23679, −64.361611 22.178 21.031 229.1 ± 9.0 2.0 Ph

A152, page 24 of 24